The Bubble and the Media

Abstract

Most of the research in comparative corporate governance has, implicitly or explicitly, taken the US as benchmark. But recent corporate scandals, such as Enron and Worldcom, question this alleged supremacy. What went wrong? In this chapter we claim that the problem is not just lack of appropriate disclosure or legislation, but a more fundamental one: deficient incentives for the media to expose poor governance practices. We argue that corporate reporters have strong incentives to enter into a quid pro quo relationship with their sources, where they receive private information in exchange for a positive spin on companies' news. Since the value of this relationship is higher during booms, so, too, will the pro-company bias. We find evidence in support of this hypothesis by looking at Harvard Business School case studies.

Full text

Forthcoming in Corporate Governance and Capital Flows in a Global Economy,
Cornelius, P. and B. Kogut (eds.), (New York: Oxford University Press, 2002).
The Bubble and the Media
Alexander Dyck
Harvard Business School
Luigi Zingales
University of Chicago
November 15, 2002
Abstract
Most of the research in comparative corporate governance has, implicitly or explicitly, taken the
US as benchmark. But recent corporate scandals, such as Enron and Worldcom, question this
alleged supremacy. What went wrong? In this chapter we claim that the problem is not just lack
of appropriate disclosure or legislation, but a more fundamental one: deficient incentives for the
media to expose poor governance practices. We argue that corporate reporters have strong
incentives to enter into a quid pro quo relationship with their sources, where they receive private
information in exchange for a positive spin on companies’ news. Since the value of this
relationship is higher during booms, so, too, will the pro-company bias. We find evidence in
support of this hypothesis by looking at Harvard Business School case studies.



We thank Mark Bradshaw, Bruce Kogut, David Moss, Forest Reinhardt and Julio Rotemberg for helpful
discussions. We also gratefully acknowledge financial support from the Division of Research, Harvard
Business School, the Center for Research on Security Prices and the George Stigler Center at the University
of Chicago.
1

Nothing but a newspaper can drop the same thought into a thousand minds at the same moment. A
newspaper is an adviser that does not require to be sought, but that comes of its own accord and talks to
you briefly every day of the common weal, without distracting you from your private affairs.
De Toqueville, Democracy in America, Vol II, Section II, Chapter VI
Introduction
A growing body of literature (eg, La Porta et al, 1997) compares corporate governance
systems across countries by comparing the set of rules and regulations in these countries. Almost
invariably the United States comes atop of such rankings. In fact, the United States is often
considered the champion of the shareholders’ value model and the example other countries
should follow. At least, this was true until recent corporate scandals and the ensuing bankruptcy
of corporate giants such as Enron and Worldcom. How could the billions in transactions between
Enron and its subsidiaries be hidden? How could the false accounting charges of Worldcom not
be revealed by internal and external auditors? Is the system fundamentally flawed? Or does it
simply lack appropriate legal and regulatory instruments?
In this chapter we advance a different explanation. While transparency can and should be
improved, we do not think that the lack of proper rules is the fundamental problem. Many of the
improprieties could have been discovered even under the current disclosure system. Why, then,
were they not discovered?
To answer this question we delve deeper into the economics of information collection
and dissemination. A simple, but thus far ignored, point is that the incentives to uncover negative
information are much smaller during stock market booms, especially those periods of stock
market euphoria often labeled as bubbles. The reasoning goes as follows. The primary source of
information collection and aggregation in any financial markets is clearly the speculators. After
all, one of the advantages of having a stock market is that it motivates people to collect
information, because they can personally gain from that information. During a phase of stock
market euphoria, however, the incentives to collect negative information are very limited. In
2

order to profit from negative news a speculator has to take a short position. Short positions,
however, are very dangerous during a phase of euphoria. Waiting for the bad news to appear and
be incorporated into prices, a speculator has to factor in a great deal of what De Long et al (1990)
calls noise trader risk: the risk of the public becoming even more enthusiastic about the stock and
driving the price up. Hence, during stock market bubbles short sellers are unlikely to search for
negative information.
Even if speculators have little incentive to uncover bad information, why don’t other
sources reveal them? The problems with equity analysts are well known and are discussed in the
chapter by John Coffee in this volume. The question we focus upon here is, why don’t the media
have an incentive to uncover bad news? After all, journalists do not have to take the risk of short
selling a stock. Furthermore, they seem to have a career-concern reason to uncover negative
news. A scoop will enhance their reputation, increasing their lifetime earnings. And at first glance
there appear to be few conflicting incentives.
This reasoning, however, is too simplistic. It ignores the subtle incentives behind
information production, collection, and dissemination. In this paper we focus on this aspect. We
argue that both the production of information by the company and the dissemination of this
information by the media is seriously affected during a stock market bubble.
In a stock market bubble, prices are driven above their fundamental values. But
astronomical multiples can be reconciled with standard valuation formulas only thanks to inflated
expectations about future growth. In fact, stock market bubbles are generally associated with talk
about a “new era.” As a consequence, during a bubble investors (and thus companies) pay much
more attention to their growth forecast and thus to news about it. Illustrative accounting studies
(eg, Bradshaw and Sloan 2002) reveal that the responsiveness of stock prices to earnings
increased more than threefold during the latest period of euphoria of the 1990s. Companies’
incentives to spin news positively and to aggressively challenge bad news are therefore greatest
3

during a bubble. Unfortunately, during a period of high valuation journalists are also particularly
willing to buy into that positive spin.
We argue that a positive spin in corporate reporting arises as a result of a quid pro quo
relationship between companies and journalists. An important asset in a journalist’s professional
portfolio is the privileged sources of information he or she has access to. After all, the Watergate
scandal would have never exploded were it not for a “deep throat” tipping Woodward and
Bernstein in the right direction. One former journalist described the situation to us thus, “When I
started I thought the client was the public, but I soon learned that in practice my client is the
source.”
1
But how do journalists maintain access to these sources? How do they reward them?
We distinguish between two possibilities. One is that the informed insider has an interest
in the diffusion of information per se. For example, in the Watergate case, Richard Nixon’s
adversaries had political reasons to leak information. They did not need to be rewarded: the
diffusion of information was their own reward. The second scenario is a quid pro quo between the
source and the journalist. The source repeatedly reveals valuable information to the journalist in
exchange for a positive spin on the news being revealed. The first case is more frequent in
environments, like the political one, where there are open conflicts of interest. This is relatively
rare in the case of corporations, with the exception of contested takeovers or internal fights to
succeed a failing CEO. In general, however, all corporate insiders have a strong vested interest in
a higher stock price and, hence, in leaking only positive news. For companies, then, the quid pro
quo scenario appears more likely. The distortionary effects of this quid pro quo have been
recognized by the Securities and Exchange Commission (SEC) when it forced equal access to
companies’ conference calls.
Since companies’ valuation are particularly sensitive to news during stock market
bubbles, corporate insiders will be particularly careful in selecting their privileged journalist
sources during these periods. Stated differently, during good times insiders have an extra
4

incentive to get the good news out and to limit bad news. On the other side, journalists will find it
particularly valuable to be in the good graces of insiders of a glamorous stock. Hence, they will
be particularly careful portraying such stocks in a positive way, for fear of losing a valuable
source in the future.
Such incentives change in a downturn. First of all, companies might have an interest in
revealing bad information when other companies are not doing so well; this type of cycle is
emphasized in Rajan (1994). Second, during a downturn the valuation of a stock depends more on
its liquidation value, than on its future growth, making it less sensitive to news. A point, again,
that seems to be borne out in accounting studies on the responsiveness of stock prices to earnings.
Finally, in downturns conflicts inside a company are more likely to arise, leading to leaks in
information. Hence, production of negative information becomes more abundant (possibly
excessively abundant) during a market downturn.
We provide some indirect evidence of cyclicality in business reporting consistent with
our quid pro quo hypothesis by looking at the composition of Harvard Business School cases,
which are a widely used source of information about businesses. There are two types of Harvard
cases: field cases, which benefit from access to companies’ internal data sources, and so-called
library cases, which are compiled using only public sources. Field cases are based in part on
private information provided by the company to the Harvard professor who is writing the case.
The explicit quid pro quo for this access is that the case needs to be approved by the company
before being released.
We expect greater use of field-based cases during expansionary periods. Companies are
expected to be more willing to share the data when the picture the case writer will portray is a
positive one, that is, during expansionary phases of the business cycle. In downturns, companies
are less willing to share their information because the picture the case writer will portray is likely
to be more negative.
5

We find this pattern of cyclicality to be true. As suggested by the raw data presented in
Figure 1, in years with strong market returns, such as those we have seen from 1995–2000, there
is a high reliance on field-based cases. Market downturns, such as those in 1991, 1995, and 2001,
dramatically reverse such trends—in 2001 alone, for example, we see a 62 percent increase in
public-source cases. We document this more systematically in the text. While indirect, this
evidence suggests a relationship between the type of information used in business coverage and
stock market performance.
Figure 1 - Are incentives in uncovering bad news anticyclical?
The proportion of field-based HBS Cases and market returns
(1986-2001)
-0.15
-0.1
-0.05
0
0.05
0.1
0.15
0.2
0.25
0.3
0.35
19
8
6
19
8
7
19
8
8
19
8
9
19
9
0
19
9
1
19
9
2
19
9
3
19
9
4
19
9
5
19
9
6
19
9
7
19
9
8
19
9
9
20
0
0
20
0
1
year
real market return
0.65
0.7
0.75
0.8
0.85
0.9
Proportion of field-
based cases
Real market return Proportion of field-based cases
Can this cyclical bias in reporting explain dramatic failures in corporate governance such
as Enron? We think so. In previous research (Dyck and Zingales 2002; forthcoming), we have
shown that the media play a fundamental role in any corporate governance system by imposing
large reputational costs on managers and directors of firms that behave against societal norms.
That the media has this power, however, does not imply they use it efficiently or effectively. In
fact, in this chapter we argue that they do not. This cyclical bias in reporting weakens the outside
monitoring of companies during booms, creating the scope for egregious abuses such as Enron.
6

This far, we have treated the bubble as an exogenous phenomenon. There is, however, a
possible feedback loop between the reluctance of the media to release bad information and a
stock market euphoria. To be sustained (at least temporarily), stock market euphoria needs to
involve a large segment of the investing population. But where do these people gather
information to form their opinions on stock prices? The media, of course. Thus, a reluctance by
the media to diffuse negative information may fuel investors’ biased expectations, sustaining the
bubble a little bit longer. In fact, one can easily conceive of a scenario in which a period of
sustained increases in stock prices due to changes in the fundamentals decimates the number of
short sellers, leaving the market dependent on the media to form a correct opinion. But the media
have few incentives to portray any other than a positive perspective, fueling the bubble. In fact,
all of the bubbles followed periods of sharp increase in stock prices due to fundamental reasons
(Garber 2000).
The idea that outside monitoring is reduced during an asset bubble is not new. More than
a century ago, Walter Bagehot (1873) wrote: “The good times too of high price almost always
engender much fraud. All people are most credulous when they are most happy; and when much
money has just been made, when some people are really making it, when most people think they
are making it, there is a happy opportunity for ingenious mendacity.” Bagehot’s view figures
prominently in the accounts of euphorias and panics offered by Galbraith,
2
Kindleburger
(1989:90),
3
Shiller (2000), and Coffee (this volume). These authors, however, assume that
investors are completely irrational, or that they at least become so during moments of euphoria. In
contrast, our argument does not rely on investors being irrational. Most importantly, we focus on
the crucial channel through which euphoria and panics are spread: the media. We show that in
reporting corporate news media tend to be biased and this bias is accentuated during booms.
The rest of the chapter proceeds as follows. The first section uses the Enron example to
illustrate that much of failure in corporate governance could have been uncovered if the media
7

had chosen to pursue this information actively, and that they deliberately chose not to. The
second section gives an analysis of the incentives behind the production, revelation, and
dissemination of companies’ information. The third section provides some empirical evidence of
the cyclicality of news spinning by looking at the composition of Harvard Business School cases.
In the fourth section we discuss the implication of this cyclicality for corporate governance.
Conclusions follow.
An Illustration of the Argument: Enron and the Media
It is interesting to analyze the role played by the media in the demise of Enron. On the
one hand, a deputy managing editor of the Wall Street Journal boasted that "without the stories
that Smith and Emshwiller (two Wall Street Journal reporters) wrote, Enron would have gotten
on fine. There is no evidence that it would have collapsed."
4
On the other hand, Sherman (2002)
notes that “to excavate back issues of magazines like Forbes, Fortune, Worth, Business 2.0, and
Red Herring is to enter a parallel universe of cheerleading and obsequiousness, a universe where
applause obliterated skepticism.” Are the media the savior or are they part of the problem?
To answer this question let’s look at some facts, as reported by Sherman (2002). The first
inquiry into Enron’s accounting came from an article by Jonathan Weil, then a reporter for the
Texas edition of the Wall Street Journal. In a July 2000 piece, Weil writes “what many investors
may not realize is that much of these companies' recent profits constitute unrealized, noncash
gains. Frequently, these profits depend on assumptions and estimates about future market factors,
the details of which the companies do not provide, and which time may prove wrong.” Weil's
piece was never published in the national edition of the Wall Street Journal, but it appeared on
the Dow Jones newswire, where it attracted the attention of James Chanos, a hedge-fund
manager, who began to scrutinize the company's financial statements. There he discovered cagey
references to "related party" transactions involving Enron's senior officers and massive insider
8

selling. It was enough to induce him to sell short the stock in November 2000. Short sellers,
however, only profit when the negative information becomes public. Thus, Chanos tipped off a
reporter at Fortune, Bethany McLean, who in March published a story entitled, "Is Enron
Overpriced?", in which she questioned how, exactly, Enron made its money. Another short seller
tipped off Peter Eavis, of TheStreet.com, who in an article on May 9, 2001 started to mention
shady "related entities."
Yet, it took Skilling’s surprise resignation in August for three Wall Street Journal
reporters (Friedland, Smith, and Emshwiller) to start an investigation of Enron's financial
statements. They quickly realized that "things weren't adding up at Enron" (Sherman 2002). Most
importantly, “sources close to Enron began to furnish the Journal with documents” (Sherman
2002).
But it was only after Enron announced a $618 million third-quarter loss that the Wall
Street Journal (October 2001) identified the link between earnings and the shady partnerships.
The Journal was able to link Enron’s reduction in shareholder equity to the CFO’s mysterious
partnerships. Only at this point did other media and the SEC start to investigate.
This brief account of Enron’s demise illustrates several important points about the role of
the media. First, while many transactions were concealed, there was enough public information
available to raise serious doubt about the credibility of Enron’s earnings. As the former editor of
the Financial Times, Richard Lambert, stated: “the Annual Report for 2000 should have raised all
kind of questions about the group’s cash flow…about the length and complexity of the footnotes
to the account, often a warning that things are not what they appear.”
5
This opinion was reiterated
by the CEO of Pearson, the group owning the Financial Times, in a recent interview to the Royal
Society of Arts Journal: “sometimes I do think that the business press—and I include the FT in
this—has not worked hard enough to ferret out these stories.”
9

The second important lesson is that the media as a whole, instead of scrutinizing Enron’s
accounts, acted as cheerleader all the way to the end. According to Jonathan Weil of the Wall
Street Journal, financial journalists “outsourced their critical thinking skills to Wall Street
analysts, who are not independent and, by definition, were employed to do nothing but spin
positive company news in order to sell stock. There was hardly a Wall Street analyst covering the
stock whose firm was not getting sprinkled with cash in some form or another by Enron.” The
day before Jeffrey Skilling resigned, Business 2.0 featured his photo on the cover with the titles
"The Revolution LIVES." In its September 2001 issue, Red Herring insisted: "Forget about
Microsoft. America's most successful, revered, feared - and even hated - company is no longer a
band of millionaire geeks from Redmond, Washington, but a cabal of cowboy/traders from
Houston: Enron."
6
The third point is that such behavior is not an accident, but the result of reporters’
incentives. Any attempt to report negative information or simply to question the existing
optimistic consensus incurs constant harassment from the target company. One UBS
PaineWebber analyst was fired three hours after issuing a warning about financial deterioration at
Enron, followed by a retraction of the negative statement and UBS PaineWebber’s issuance of an
optimistic outlook on Enron’s future.
7
During an investor conference call, a caller who criticized
Enron’s delays in releasing financial information was labeled an “asshole” by Skilling.
8
Similarly, Fortune’s journalist McLean was labeled “unethical” by Skilling, who hung up on her.
Furthermore, the chairman of Enron, Ken Lay, called Fortune’s managing editor Rik Kirkland,
implying that McLean’s piece should be cut.
9
The last important point is that the incentives to report bad information about a company
change dramatically as the company’s stock price deteriorates. One could simply appeal to the
journalists’ herding mentality, but we think there is more to it than that. As Sherman’s (2002)
account of the Enron story demonstrates, as the stock price falls, journalists start to have access to
10

more negative information. Short sellers start feeding them and so do company’s insiders, who
hope to benefit from a turnaround of the company. These sources were not available when the
stock price was booming. When a company’s fortunes deteriorate not only do journalists have
access to more negative news, they also have less reasons to hide them. The possibility of a
demise of the company reduces the value of an ongoing relationship, increasing a journalists’
willingness to report negative information.
The economics of media reporting is, therefore, not as straightforward as economic
models implicitly assume. Thus, to understand the effect of a stock market bubble on the quality
and type of media reporting, we need to closely study the incentives to generate, reveal, and
disseminate information. This is the task to which we now turn.
Media Incentives and Governance
The economics of media coverage
To understand the economics behind news media we need to start appreciating their role.
The main role of news media is to filter and aggregate information, repackaging it with some
entertainment value. For example, the dry and prolix prose of police reports is transformed into a
thrilling story about the latest murders. As in any other markets, the type of information produced
and its quantity depend on supply and demand considerations. While these considerations are
similar across topics, some issues are unique to corporate news. Since this is the area in which we
are interested, we will confine our reasoning to this case.
The positive externality of media for corporate governance
The presence of an active press can increase expected penalties for improper behavior
through at least two channels. In the conventional reputation story (eg, Fama 1980), managers’
wages in the future depend on shareholders’ and future employers’ beliefs about whether the
11

managers will attend to their interests in those situations where they cannot be monitored.
Managers, however, care about their reputation not only vis-à-vis future employers, but also vis-
à-vis society at large. Commenting on the recent legal reforms in US corporate governance,
Robert Mills, managing director of UBS Warburg, recently argued that for him a tougher
punishment than the threat of jail time would be explaining the story about it in the Wall Street
Journal to his son.
10
The strength of reputation as a corrective device, however, depends upon the extent to
which information about past behavior is disseminated. As the opening quote from De Toqueville
suggests, and as we develop in Dyck and Zingales (2002), the media play a pivotal role in
communicating information about managers to the public. For instance, in Dyck and Zingales
(forthcoming) we show that the diffusion of information by the press affects the amount of
corporate value that insiders appropriate for themselves, the so-called private benefits of control.
Across 39 countries, a higher rate of diffusion by the press is associated with lower private
benefits of control, and this influence persisted even controlling for differences in laws and legal
regimes.
11
This evidence is consistent with many statements by executives. As William Browder,
managing director of Hermitage Capital, the largest public equity fund in Russia claims, “The
court of public opinion is much more effective than the Russian legal system and much fairer.”
12
The economics of information production
Arguing that the media has a positive externality for corporate governance does not imply
that the news media will provide the optimal amount of coverage to address all governance
problems. To answer the question of why the optimal coverage is not likely to be provided, we
need to focus on the incentives to uncover bad information. Consider, first, the supply side.
Collecting information about companies is very costly. Digging through annual reports,
questioning the validity of different accounting practices, and so on require both time and
expertise. Most important, it is a very risky activity. Without a lead, a random investigation will
12

turn out to be fruitless most of the time. Unless he or she uncovers a major fraud, a journalist who
spends a lot of time digging through annual reports has little to show for it. According to Sherman
(2002), it took two months for Jonathan Weil to prepare the Enron article, and this article did not
even make it to the national edition of the Journal.
A much easier (and safer) alternative for a journalist is to rely on companies’ press
releases. In fact, the stated goal of these releases is precisely to help journalists quickly absorb a
great quantity of facts. Of course, these releases are not unbiased: they try to spin the story in the
direction that is most favorable to the company. While expert reporters can easily see through the
spin, a complete undoing of the built-in bias requires time and effort. Whether the market will
adequately compensate journalists for this time and effort depends upon the structure of the
demand for information, a question we will return to soon.
From a reporter’s point of view the real downside of relying on press releases is the lack
of any competitive edge. A reporter wants to differentiate his or her product, wants to report
critical information before other reporters. What constitutes critical information, however,
depends upon what other reporters write. It is a game similar to Keynes’ famous beauty contest,
where everybody tries to guess what everybody else prefers. In this game the real competitive
edge is personal contacts. Sources inside or very close to the company that will tip on major
coming news ahead of the rest of the crowd.
These contacts are extremely valuable for reporters, who cultivate them actively. But they
do not come for free. News sources generally have a reason to leak information. They want to
scare off a competitor, reassure shareholders, prevent a board coup, and so on. In the political
arena, it is the fierce competition between opposing parties with different agendas that generate a
relatively unbiased set of information sources. But in the corporate world, such competition is
lacking. Except for short sellers, all informed parties have a vested interest in a high stock
valuation: managers, who hold stock options; employees, whose jobs depend on the company
13

doing well and whose retirement accounts depend on its stock doing well; and analysts, whose
fortunes are very often linked to the success of the stock they analyze. If all the sources have an
interest in a positive spin, the news coming from them will be clearly biased.
Once again, a reporter could potentially undo this bias, but in addition to the problems
discussed above, he or she faces an additional constraint: the need to reward sources. And since
all these sources have a vested interested in a high stock price, the way to reward them is to spin
the story in a positive direction. The last thing reporters want to do is develop a reputation for
writing negative articles—they will find it extremely difficult to develop and maintain their own
sources and they will face constant harassment in doing their job (as happened to the journalists
writing about Enron). A possible remedy for a media outlet is not to disclose the name of writers
(as the Economist does). Hiding their identity makes it easier for reporters to write negative
stories about companies. But even this device is imperfect. Most companies know which
journalists are assigned to their case and they can easily infer who is writing. Hence, it is
impossible to remove completely reporters’ positive bias toward the companies they cover.
Thus far, we have also argued that producing negative news is much more costly than
producing positive news about a company. This does not necessarily mean that in equilibrium
less bad news will be produced—it depends on the price attached to one type of news versus the
other. Will negative news be rewarded more than positive news? Will the premium be sufficient
to reward the additional costs negative news involves?
We doubt that. While a scoop benefits a reporter, the benefits she receives are not
different if the scoop reveals bad information or good information about a company. The costs
the reporter receives, however, do depend on the type of scoop. . In fact, negative scoops are
more dangerous, because in some countries they expose a reporter to the risk of being sued for
defamation by the company. Even if the article is correct, the cost of the trial can easily break a
reporter both emotionally and financially. By contrast, no such a risk exists for a positive scoop.
14

Even if the scoop turns out to be false, the reporter is not sued by the myriads of investors who
bought the stock on the basis of the false scoop. Hence, positive scoops are more rewarding for a
reporter than negative ones.
But even if negative scoops were rewarded more, the incentives to seek such a scoop
could remain excessively low. The risk involved in a negative scoop is such that a risk-averse
reporter might not seek one even in the presence of a premium.
Cyclicality in the media’s bias
Thus far, we have simply described how the economics of information production lead
the media to report, on average, a positive image of a company. We have said nothing about
possible variations in the magnitude of this bias over time. Now we will argue that this pro-
company bias is stronger during a boom and weaker, to the point of being reversed, during a
downturn.
The first reason for the bias discussed is the cyclicality in the availability of different
sources of information. As argued above, speculators are the primary source of information
collection and aggregation in any financial market because they can personally gain from that
information. During a phase of stock market euphoria, however, the incentives to collect negative
information are very limited because of the dangers of noise trader risk: the risk of the public
becoming even more enthusiastic about the stock and driving the price up. During the dot-com
frenzy Lamont and Thaler (2001) document deviations from the fundamental value of more than
30 percent even when it was possible to perfectly hedge a short position (as in the case of Palm
Pilot and 3Com). One can only imagine what happens when such a perfect hedge does not exist.
Hence, during stock market bubbles short sellers are unlikely to search for negative information
or to tip off to the media. Similarly, no insiders want to rock the boat when the company is doing
well and it trades at high multiples. As Sherman’s (2002) account describes, Enron insiders
started to leak internal information only after the resignation of Skilling. Before, nobody dared to.
15

The second reason for this cyclicality is the asymmetry in the value of the quid pro quo
relationship over the business cycle. When a company is doing well it has a lot of positive news
to leak to the media. Hence, antagonizing a successful company with a negative report can be
very costly for a journalist. The “uncooperative” reporter will be denied access to companies’
data, a loss to the reporter of a valuable source of privileged information. By contrast, in a
downturn a company will have very little information it wants to share with reporters. Therefore,
antagonizing it with a negative report and being cut off from access to company data will not be
so costly. If, in addition, the company is on the verge of bankruptcy, the scope of the quid pro quo
relationship with reporters is reduced and the temptation for the reporter to deviate and report bad
information increases.
The third reason why negative reports about companies are anticyclical is that the effort
companies put into preventing them is procyclical. During a boom phase, a lot of the value of a
stock depends on a company’s prospects for future growth. And these prospects are highly
influenced by the information released by the media. In contrast, in a market downturn, the value
of a stock is more driven by the value of its asset in place, which is more tangible and thus less
affected by any story reported in the media. Hence, the effort companies might devote to spinning
news for the press is going to be much more intense during a boom phase than during a
downturn.
13
Similarly, it is more costly for a reporter to antagonize a company when he or she is the
only one doing it, than when many others are doing so. Questioning the integrity of a company’s
numbers when the company is doing well is very dangerous. A single pundit or reporter can be
easily harassed or even sued, since the company can hope, with this strategy, to prevent others
from following the first’s example. But when a company is openly questioned by multiple
sources, the aggressive strategy becomes self defeating and each reporter runs little risk of being
harassed or sued. This asymmetry can, by itself, generate a herding behavior among reporters.
16

Illustrative of the strong forces to suppress “bad” information is the reception that Alan
Greenspan received for the rather mild criticism he offered on December 6, 1996 when he asked,
"How do we know when irrational exuberance has unduly inflated asset values, which then
become subject to unexpected and prolonged contractions…? ''
14
Immediately, he was criticized
as being not only out of touch with the economy, but for jeopardizing the economic boom. That
Greenspan, one of the most independent of market commentators and who had a well established
reputation, was subject to severe criticism hints at the ferocity with which bad news is greeted by
those with less independence and reputation, such as the equity analysts and newspaper reporters
described above in the Enron example.
This phenomenon is not unique to the 1990s. Paul Warburg (a banker) and Roger Babson
(a statistician) were exposed to public condemnation for their early criticism of speculation in
1929 (Galbraith 1990). Galbraith himself received death threats for his mild criticism of the
speculative buildup in 1955: “The postman each morning staggered in with a load of letters
condemning my comments, the most extreme threatening what the CIA was later to call executive
action, the mildest saying that prayers were being offered for my richly deserved demise”
(Galbraith 1990:9). And in 1986 the New York Times refused to publish an article it
commissioned to him when he concluded that a crash was inevitable.
A more elaborate version of such bias coming from the consumers of business news is
suggested by Mullanaithan and Shleifer (2002). They assume that readers believe more articles
that confirm prior articles, while they discount the others. With such behavior it pays for reporters
to follow the herd. During boom periods, they will accentuate positive news, while in downturns
they will emphasize negative ones. While this behavior is sufficient to generate cyclicality in
media reporting, it cannot fully account for some of the evidence we will discuss later.
Extensions
17

In all the above discussions we have implicitly assumed that individual reporters have
full discretion in their choices. In fact, which topics are investigated and how much time is
dedicated to single investigations is decided in part by media headquarters. Much of the analysis,
however, carries through at that level as well. In fact, at that level the pressure to please
companies can be even more severe, since companies’ advertising is a major source of revenue.
We have also ignored the possibility of outright corruption, where companies pay
reporters or newspapers to spin their versions of the facts and/or suppress bad information. We
have done so because in the United States (but not necessarily in other countries) this is extremely
rare. Kindleburger (1990), under the heading, “venal journalism,” cites numerous examples of the
press being bought by speculators during the South Sea Bubble in the United Kingdom and on the
Continent in the 19
th
century, but suggests that the United States has been much less open to such
problems.
This is not to say that there have been no charges. Journalist Phillip Longman contends
that in the US, “business publications, especially those celebrating the boom, were growing fat
from dot-com ads…at any publication, there is, of course, a tension between the need to please
advertisers and the need to please readers. … at business publications in the 1990s, it [this
tension] was resolved in a manner that favored advertisers—and worse, advertisers, who, it has
turned out, were selling shoddy products.”
15
In the case of Enron, questions have been raised
about the conflicts many journalists faced as a result of direct or indirect payments from Enron
through speeches and positions on Enron’s advisory board. Josh Lipton (2002) reports Andrew
Sullivan’s challenge to journalists: "Exactly how many pundits have been on Enron's payroll?
How many of them have disclosed that fact in their relevant publications? How much was each
paid?" In this respect, conflicts may be at least as severe for academics writing about business
issues, through their multiple contacts with companies through consulting engagements and board
and advisory positions.
18

Of course, allowing for that possibility will only strengthen our reasoning, both in terms
of the average bias and in terms of its time series variation. Companies will have both more
resources with which to bribe and more benefits from a positive spin in a boom than in a
downturn, hence they will end up bribing reporters more.
Empirical Evidence on the ‘Spin’ of Business Coverage and Stock Market Cycles
What evidence is there that companies try to spin information to make their performance
look stronger? And what evidence is there that the media is more inclined to take or exaggerate
this spin during booms than downturns?
Companies spinning information
Existing research on press releases of company financial information shows efforts by
companies to spin. It is often the case when companies reveal information that they also reveal
information from a previous year to facilitate comparisons. Reflective of strategic spinning of
information, Schrand and Walther (2000) report that companies systematically tilt these
comparisons to make current performance look stronger. They are much more likely to remind
readers of extenuating reasons for strong performance in previous years than to remind readers of
extenuating circumstances for poor information.
Bradshaw and Sloan (2002) provide complementary evidence based on what firms’
define and emphasize in their press releases surrounding mandatory filing of earnings. The
traditional definition of earnings is that required by generally accepted accounting principles
(GAAP). But there are alternatives, variously called pro forma earnings, operating earnings, or,
most commonly, “street earnings.”
16
These alternatives are suspect, at least according to the
chief accountant of the SEC, as the Wall Street Journal reported: “Mr Turner said it appears as if
some companies are intentionally trying to ‘spin investors’ by issuing news releases highlighting
19

pro forma earnings, which tend to omit items that would reduce reported earnings. Mr Turner
jokingly called such earnings figures ‘everything but the bad stuff.’"
17
Bradshaw and Sloan (2002) report that looking at the average of all US equities, street
earnings have exceeded GAAP earnings every year since 1987. And again, consistent with
spinning, companies have been in the lead in defining and emphasizing these earnings, with the
typical press release announcement emphasizing street earnings earlier than GAAP earnings
when this interpretation would put the company in a better light. In 1998–1999, for instance, a
buoyant market, this spin was put on the numbers 43.5 percent of the time.18
1
Business coverage open to spin and stock market cycles
Now we turn to evidence that business coverage is more open to spin in boom times than
downturns. Our empirical evidence comes from another source of business information—case
studies prepared by Professors at Harvard Business School. This is not the traditional business
press; in fact the business school has as a disclaimer on all cases: “Cases are not intended to serve
as endorsements, sources of primary data, or illustrations of effective or ineffective
management.” Nonetheless, the case studies are an important source of information on business:
Harvard Business School case studies are sold all over the world, they form the basis of the
curriculum at many business schools, and they are used as sources in many news articles.
We focus our analysis on the two most common types of Harvard cases: field cases,
which benefit from access to companies’ internal data sources, and the so-called library cases,
which are compiled using only public sources. Field cases are the most typical type of case study.
They are based in part on private information provided by the company to the Harvard professor
who is writing the case. The explicit quid pro quo is that field cases require approval by the
company before release. In fact, the most objective way of classifying a case study as a field or
library case is whether the case writer has felt compelled to get a “green card” for the case; that is,
20

a green document that a company representative has to sign to release the case study that contains
private information.
That field cases, which are subject to companies' approval, might be more influenced by
the companies involved than are cases from public sources, is illustrated by the following
episode. "I learned that Enron was upset with my public-source case on the conflict surrounding
the company's investment in India," recalls Harvard Business School professor Louis Wells.
"After the second time the case was taught, someone from the administration approached me, told
me of the company's concerns, and asked if anything could be done about it. Another faculty
member was, I was told, writing a field-based case on the same subject. It was suggested that I
might consider teaching the more rich field case, if it fit my teaching objectives. Meanwhile, I
sent my public-source case to Enron for comment. In the end, I removed the public-source case
from the system and adopted a shortened version of the field case, which was indeed richer in
information and enabled me to accomplish the original teaching objectives.”
19
The advantage of looking at Harvard Business School cases is that it is possible to
identify when the quid pro quo takes place. By looking at the composition of case studies over
time we can see if the proportion open to spin is influenced by the market. If our hypothesis about
cyclical behavior is correct, we expect greater use of field-based cases during expansionary
periods. When the case to be written is meant to portrait a positive image of the company ( ie,
during expansionary phases of the business cycle), the implicit price to be paid for access to
privileged information is lower. Hence, companies will be more willing to share their information
and Harvard Business School professors will be more willing to engage in the quid pro quo with a
company. As one of the authors attests, during the recent stock market boom there were more
unsolicited offers by companies to support case studies than in the period before the boom. In
21

downturns, unsolicited proposals for case studies are fewer; companies are less willing to share
their information as the picture the case writer will portray is likely to be more negative.
20
Our test of the theory is a simple one, does the proportion of field-based cases move
positively with the stock market? Empirically, we regress the proportion of field-based cases
against a measure of market returns (and a time trend to account for any changes in the approach
to collecting information about companies over time at Harvard Business School). To test this we
collected data from Harvard Business School on the number of field and library cases published
each year since 1970. With this information, we calculated the proportion of field-based cases in
any year.
The summary statistics of our sample are provided in Table 1. The typical case study is
field-based, with the average proportion of field-based cases equal to 82 percent with significant
variation, from 69 to 94 percent. Visual inspection of the data also shows a trend to reduced
reliance on field cases over these 30 years, so we will account for this possibility in our empirical
analysis. The remainder of the table simply summarizes well-known trends in real market returns
and in real GDP over this time period, with an average real market return of 4.9 percent and an
average increase in real GDP of 3.1 percent. As the Business School has grown, so has the
number of case studies, with an increase over time from 56 case studies in 1971 to 406 case
studies in 2001.
Table 2 confirms what was suggested by Figure1—business coverage that relies on
sources prone to spin is cyclical. The positive and significant coefficients on market returns in
columns 1 and 3 show that over the whole sample (1971–2001) and, particularly, in the latter half
of the sample (1986–2001) field-based cases are sensitive to market returns. The results suggest
that a one standard deviation improvement in market return brings forth a 1.6 percent increase in
field-based cases in general, and that this sensitivity is increasing, with an estimated 3.3
percentage point increase in field-based cases for a one standard deviation increase in market
22

returns during the mid-1980s and 1990s. The effect of stock market downturns appears
particularly important. As an illustration, in the stock market downturn of 2001, field-based cases
account for just 75 percent of all cases, a full 10 percent drop from the good times of 2000. The
same was true in the last significant market downturn in 1991, where field cases accounted for
just 68 percent of all cases, while two years earlier in better times they accounted for 82 percent.
Columns 2 and 4 show that the sensitivity is not only to market returns, but also to the
state of the macroeconomy, and that again, the sensitivity is highest in the latter half of the period.
Our regression estimates suggest that over the whole sample period (the latter half) a one standard
deviation change in the economy brings forth a 1.7 percent (5.5. percent) increase in field-based
cases.
Does this matter?
Now that we have (hopefully) convinced the reader that media are biased in their
reporting and that this pro-company bias is stronger during booms than during recessions, we can
ask the more fundamental question: does it matter?
In a frictionless world, where speculators are free to short a stock and they are not afraid
to do so, this media bias will have no serious consequence on prices. Speculators have no bias in
one direction or the other and hence they will force prices to their fundamental values.
In the real world, however, short sales are not as easy as long positions. Hence
speculators find it more difficult to correct overvaluations than undervaluations. As a result,
overvaluations can persist. Unfortunately, the described bias in media reporting does nothing to
reduce this problem. In fact, it exacerbates it. When the force of speculation is crippled by short-
sell constraints, the media, instead of convincing the public-at-large that the prices are above
23

fundamentals, will tend to feed the exuberance. In other words, while the media are probably not
responsible for the rise of bubbles, they clearly play a part in sustaining them.
Even if this cyclicality in media does not affect prices, however, it does have important
effects on corporate governance. As we show in Dyck and Zingales (forthcoming), the pressure
exerted by the media is an important component of a good corporate governance system. When
such pressure weakens, abuses are inevitable. “The press blithely accepted Enron as the epitome
of a new, post-deregulation corporate model,” stated Business Week in an unusual mea culpa,
“when it should have been much more aggressive in probing the company's opaque partnerships,
off balance sheet maneuvers, and soaring leverage." Unfortunately, our analysis suggests that this
is not an occasional lapse, but a systematic problem that emerges during stock market booms.
That media’s incentives to uncover negative information are weakened during booms
does not mean media play no role in corporate governance. As the Enron episode illustrates, the
press still played an important role in stopping abuses at Enron, and has played a very important
role in imposing penalties long before any legal penalties are introduced. The point we want to
stress here is that during booms the monitoring provided by the media is less than that provided
during normal or bad times. The fact that the weakening of media’s incentives to uncover bad
news could lead to such egregious behavior such as Enron emphasizes the importantance of the
media. After all, the quality of the legal system did not change over the stock market cycle. For
what other reason should corporate behavior have changed?
What can be done?
If this problem is indeed so severe, what can be done to attenuate it? Our analysis of
media incentives suggests one possible solution. The reason why reporters engage in quid pro quo
relationships with companies is to reduce the cost of collecting information. Thus, the higher this
cost is, the stronger these relationships will be. After all, the percentage of public information
cases is higher in accounting and finance (where case writers have access to better public sources)
24

than in entrepreneurship and strategy, where public sources are lacking. Thus, to weaken these
ties between reporters and companies, enhanced corporate disclosure can be useful. Our analysis
suggests, however, that the presence of the information in the public domain is not sufficient; the
cost of gathering is also very important. The higher this cost, the bigger will be the incentive for a
journalist to skip this cost and rely on direct company sources. Consider, for instance, the
accounting treatment of stock options. Many have argued that disclosure in the footnotes is
sufficient (ie, what is required under current accounting provisions), since sophisticated investors
can calculate the implied costs of options and restate financials. Our perspective suggests
otherwise. This indirect form of disclosure raises the costs for journalists to use this information
to uncover or communicate corporate misdeeds.
Conclusions
In this chapter we advance a new explanation for the severe lapses in corporate
governance experienced by US companies in the last few years. This explanation focuses on the
role of the media, an ignored, but we think important, institution in determining governance
outcomes. We argue that during stock market booms the healthy pressure exerted by the press on
companies is weakened, because reporters find it more convenient to buy into companies’ spin.
We provide some indirect evidence of this effect by looking at the percentage of Harvard
Business School cases that are field-based, that is, that rely on companies’ internal sources.
Consistent with our quid pro quo theory, this percentage increases during booms and drops during
recessions.
We suggest that greater availability of ready-to-use public information can reduce
reporters’ incentive to enter into quid pro quo relationships, maintaining greater independence of
the press even during booms.
25

Table 1
Summary Statistics
Variable Mean Median Standard
Deviatio
n
Min. Max. Number
of Obs.
Percentage of field-based
cases
0.816 0.816 0.052 0.686 0.935 31
Real market return 0.049 0.070 0.17 -0.367 0.35 31
Real GDP growth 0.031 0.035 0.022 -0.020 -0.073 31
Number of Harvard Business
School cases
245 276 108 56 406 31
Table 2
Is There a Cyclical ‘Spin’ in Business Coverage?
The dependent variable is the percentage of Harvard Business School case studies that are field-
b
ased by
publication year (1971–2001). This is defined as the ratio of the number of field-based case studies (field-
based cases require written approval from firms before use) to the total number of field-
b
ased and library cases
(library cases do not require firm approval before use) in a given publication year. (Source: Harvard Business
School Publishing.) In column 1 and 2 we regress the percentage of field-based case studies against the real
market return and real GDP growth respectively using the whole sample period of 1971–2001 and including
controls for a time trend and a constant. In columns 3 and 4 we repeat this analysis restricting ourselves to
1986–2001, the second half of our sample period. The real market return is the percentage change in the end o
f
year Dow Jones index less the percentage change in the Consumer price index in that year. (Sources: Dow
Jones, Economic Report of the President). Real GDP growth is the percentage change in real GDP. (Source:
Economic report of the President). A positive coefficient on real market return and/or real GDP growth
indicates a sensitivity of use of field-based cases to market and macroeconomic performance.
Independent Variables Dependent Variable: Percentage of Field-Based Case
Studies by Publication Year
(1) (2) (3) (4)
Real market return 0.095* 0.193**
(0.054) (0.083)
Real GDP growth 0.775* 2.52***
(0.394) (0.538)
Time trend -0.003*** -0.002** -0.002 -0.002
(0.001) (0.001) (0.002) (0.002)
Constant 6.75*** 5.13*** 5.06 5.30
(2.05) (1.86) (4.40) (3.19)
Years covered 1971–2001 1971–2001 1986–2001 1986–2001
Number of years 31 31 16 16
Adjusted R-squared 0.18 0.20 0.22 .59
Notes: *** denotes significant at 1%; ** denotes significant at 5%;* denotes significant at 10%.
26

References
Bagehot, W. 1873. Lombard Street: A Description of the Money Market. Foreword P. L.
Bernstein. 1999. Reprint, New York: John Wiley & Sons.
Bradshaw, M. and R. Sloan. 2002. “GAAP versus The Street: An Empirical Assessment of Two
Alternative Definitions of Earnings,” Journal of Accounting Research 40(1):,41–66.
De Long, J. B., A. Shleifer A, L. H. Summers, R. J. Waldman. 1990. “Noise Trader Risk in
Financial Markets,” Journal of Political Economy 98(4):703–738.
Dyck, A. and L. Zingales. 2002. “The Corporate Governance Role of the Media.” In The Right to
Tell: The Role of Mass Media in Economic Development. Washington, DC: The World
Bank..
Dyck, A. and L. Zingales. Forthcoming. “Private Benefits of Control: An International
Comparison,” Journal of Finance.
Galbraith, J. K. 1990. A Short History of Financial Euphoria. New York: Viking Press.
Garber, P. M. 2000. Famous First Bubbles : The Fundamentals of Early Manias. Cambridge,
MA: MIT Press.
Kindleburger, C. 1989. Manias, Panics and Crashes: A History of Financial Crises. New York:
Basic Books.
La Porta, R., F. Lopez-de-Silanes, A. Shleifer, and R. Vishny. 1997. “Legal Determinants of
External Finance,” Journal of Finance 52(3):1131–1150.
Lamont, O. and R. Thaler. 2001. “Can the Market Add and Subtract? Mispricing in
Tech Stock Carve-Outs,” NBER Working Paper W8302. Cambridge, MA: National
Bureau of Economic Research.
Lipton, J. 2002. “Ethics: Enron's Helpers,” Columbia Journalism Review 40(6):13.
Mullanaithan, S. and A. Shleifer. 2002. “Media Bias,” Harvard University, mimeograph.
Rajan, R. G. 1994. “Why Bank Credit Policies Fluctuate: A Theory and Some Evidence,”
Quarterly Journal of Economics 109(2):399–441.
Schrand, C, and B. Walther. 2000. “Strategic Benchmarks in Earnings Announcements: The
Selective Disclosure of Prior-Period Earnings Components,” The Accounting Review
75(2):151–177.
Sherman, Scott. 2002. “Enron: Uncovering the Uncovered Story,” Columbia Journalism Review
40(6):22–29.
Shiller, R. 2000. Irrational Exuberance. New York: Broadway Books.
Endnotes
1
Jonathan West, interview with author, November 17, 2002.
27

2
As Galbraith argues in The Great Crash 1929 (Boston: Houghton Mifflin, 1979), "In good
times, people are relaxed, trusting and money is plentiful. ... under these circumstances the rate of
embezzlement grows, the rate of discovery falls off, and bezzle increases rapidly.”
3
“In a boom, fortunes are made, individuals wax greedy and swindlers come forward to exploit
that greed.”
4
Quoted in Sherman (2002).
5
Quoted in: Johnson, R. 2002. “City Scribblers Smarting Over Questions of Competence,” The
Daily Telegraph Friday October 11, 2002:18.
6
All quotes in this paragraph are from Sherman (2002).
7
Based on documents revealed by Representative Waxman of California in the House
Government Reform Committee and reported in: Richard Oppel Jr, R. 2002. “The Man Who Paid
the Price for Sizing up Enron,” The New York Times March 27, 2002:C1.
8
Smith, R. and J. Emshwiller. 2001. “Enron Prepares to Become Easier to Read,” Wall Street
Journal August 28, 2001:C1.
9
Quoted in Sherman (2002).
10
Wharton, Institutional Investors as a Force for Change, Reviewed November 12, 2002. Online.
http://knowledge.wharton.upenn.edu/articles.cfm?catid=1&articleid=655&homepage=yes
11
In this chapter we use differences across countries in the diffusion of the press to identify the
impact of the press. Because the press cannot be important if their reports are not read, diffusion
is clearly a rough indicator of media importance, but one of the few available in a large cross-
section of countries.
12
Dyck, A. 2002. “The Hermitage Fund: Media and Corporate Governance in Russia,” Harvard
Business School Case Study 2-703-010. Cambridge, MA: Harvard Business School.
13
On the other hand, a CEO facing a downturn might be particularly interested in protecting his
public image.
28

29
14
Remarks by Chairman Alan Greenspan at the Annual Dinner and Francis Boyer Lecture of the
American Enterprise Institute for Public Policy Research, Washington, DC, December 5, 1996.
Greenspan, A. The Challenge of Central Banking in a Democratic Society. Reviewed September
30, 2002. Online. http://www.federalreserve.gov/boarddocs/speeches/1996/19961205.htm
.
15
Longman, P., 2002, “Bad Press,” Washington Monthly 34(10):19.
16
Bradshaw and Sloan (2002:45) report that these definitions exclude supposed nonrecurring
items such as restructuring charges, merger and acquisition costs, goodwill amortization, certain
results of subsidiaries, stock-based compensation costs, and even, in the case of Amazon, interest
expense on long-term debt.
17
Wall Street Journal. 2001. “ SEC Probes 4 Firms For Possible Abuses Of Pro-Forma Results,”
Wall Street Journal June 19, 2001:C18.
18
Based (in each period) on a random sample of 200 earnings press releases where the street
numbers exceeded the GAAP numbers
.
19
Louis Wells, in an interview with the author, November 15, 2002.
20
The Mullanaithan and Shleifer (2002) herding theory does not predict any push from the supply
side. Hence, this episode, as well as Enron’s sharing of data sources in exchange for a better spin
of the case, is only consistent with the quid pro quo theory.