The Press as a Watchdog for Accounting Fraud

Gregory S. Miller
Graduate School of Business Administration
Harvard University
Boston, MA 02163
gmiller@hbs.edu

First Version: July 1, 2003

This Version: December 29, 2003

Preliminary and incomplete. All comments welcomed.

I thank Mary Barth, Sudiptu Basu, Bob Kaplan, Stu Gilson, Cristi Gleason, Pat Hopkins, Bruce Johnson, Michael
Maher, Maureen McNicholas, Greg Waymire and Joe Weber as well as workshop participants at Boston College,
Emory University, University of Iowa and the 2003 Stanford Summer Camp for comments on earlier versions of
this paper. I thank Sarah Eriksen and Anne Karshis for research assistance. I am grateful for the funding of this
research by the Harvard Business School.
 The Press as Watchdogs for Accounting Fraud

ABSTRACT

This paper investigates the press’ role as watchdogs for accounting fraud. In choosing whether
to fulfill this role, the press must trade off the benefits of providing interesting stories, and thus
increasing circulation, with the costs of identifying such stories and potentially alienating
business partners and advertisers. I use a sample of firms that the SEC has found guilty of
committing accounting fraud to investigate how the accesses these costs and benefits. I find the
press is more likely to write an article identifying accounting fraud if the firm has a high
information environment (i.e. many press articles in general, high analyst following and,
although weaker results, a large market value of equity). A rich information environment
indicates the press can obtain information at a lower cost and that a large number of potential
readers would be interested in the firm. I also find that characteristics of the fraud impact the
likelihood the press will write an article. Fraud that includes material publicly misleading
statements or management misappropriation of funds is more likely to trigger an article. A
public statement can both attract the presses attention, thus reducing identification cost, and
assure more people are interested in a follow-up to information they have used. Management
misappropriation of funds, whether theft or insider trading, is likely to make a story more
interesting to the public, and thus increase the benefits to the writer. I do not find the severity of
the violation, based on length and number of violations noted, to influence the likelihood of an
article being written. Finally, I perform a content analysis of the identified fraud articles to gain
a greater understanding of how the press undertakes the watchdog role. I find that the press
relies on many sources, but particularly on analysts, legal suits and auditor changes. While many
types of press break the initial story, the business press provides the greatest coverage. Finally,
most articles do not appear to be written by local reporters, but those that do are more likely to
provide no source for information (consistent with having performed their own investigative
reporting) and are less likely to rely on analysts in general. This provides some evidence of an
informational difference for locally based reporters.
“More crime, immorality and rascality is prevented by the fear of exposure in the newspapers
than by all the laws, moral and statute, ever devised.” – Joseph Pulitzer, 1878 as cited in
“Muckraking! The Journalism that Changed America”

1. Introduction

This paper studies the press’ role as a monitor or “watchdog” identifying accounting fraud.

Using a sample of firms that have been sanctioned by the SEC for accounting violations, I

examine the relation between firm and fraud characteristics and the violation is identified by the

press prior to a disclosure of the violation by the firm or SEC. For the subset of firms initially

identified by the press as committing fraud, I also provide content analyses of the articles to

examine the information sources the press used to identify fraud, and the types of publications

that undertake this watchdog role. The evidence provided by this study contributes to our

understanding of the press’ role as an information intermediary in the capital markets and

society.

The press has a long and storied history of “muckraking” investigative reporting in various

aspects of society. Such investigative reporting entails combining widely available and “new”

information to highlight a wrongdoing. This watchdog role is often cited as one of the most

important functions of the press (Serrin and Serrin 2002; Islam 2002; Djankov, Mcliesh, Nenova

and Shleifer 2002; Dyck and Zingales 2002). Accounting fraud presents a good opportunity for

the business press to undertake such a role. Frauds are often filled with conflict, drama and

deviant behavior, all of which are considered components of a compelling story (Jamieson and

Campbell, 2001). Compelling and entertaining stories lead to a larger reader base and a related

increase in revenues from subscriptions and advertising, creating an incentive to aggressively

fulfill the watchdog role. However, the press is also part of the market it is meant to cover. As

such, there may be incentives not to harm those markets due to relations with affiliated

1
companies, advertisers, or simply upsetting the reader-base (Herman, 2002; Jamieson and

Campbell, 2001). Further, uncovering accounting fraud is likely to be time consuming, costly

and lead to many dead ends. Thus, members of the press (both publishers and reporters) must

trade-off these costs and benefits when deciding whether and how to identify accounting fraud.

I use Accounting, Auditing and Enforcement Releases (AAER) to identify a sample of firms

that were sanctioned by the SEC for accounting fraud, allowing me to examine a sample of firms

widely believed to have engaged in accounting fraud. The details provided in AAER allow me

to capture characteristics of the fraud undertaken, such as the time period, length of the fraud and

types of infractions involved.1 My final sample consists of 263 firms that have committed a

wide range of accounting violations. I find that the press publishes articles regarding accounting

fraud prior to a public acknowledgment by the firm or SEC for 76 (29%) of the firms.

Next, I investigate how the firm information environment and fraud characteristics impact

whether the press identifies and writes about the fraud. I predict that firms with a strong

information environment are more likely to have articles written regarding their accounting fraud

due to the obvious high interest in these firms, which suggests there will be interest in a story,

and to the lower cost of investigating firms with a rich information environment. Consistent with

these predictions, I find that firms with a high analyst following and/or a large number of general

press articles are statistically more likely to have their accounting violation identified in the

press. Results on firm size (market value of equity) also provide some support for this

prediction. However, in several specifications this variable is not significant at conventional

levels.

1
As discussed in section 3.1, using AAER also has some costs as it may reduce my ability to generalize results to
the population as a whole.

2
 The press industry generates much of its income from advertising revenue. Accordingly, the

press may be less likely to be critical of firms that are currently large advertisers or have the

potential to be in the future. However, I find no evidence that the press is less likely to write

articles regarding firms in high advertising industries.

I also expect aspects of the fraud will impact whether the press publishes a fraud identifying

article. Frauds that are more egregious are likely to make for a more interesting story and be

easier to identify. Accordingly, I examine whether frauds that have a greater number of

violations cited in the AAER or last over a longer time are more likely to be identified in an

article. The evidence does not support this prediction.

Frauds that involve public misleading statements (such as a press release claiming a large

new contract) are likely to both attract the press’ attention, thus reducing search costs, and be of

interest to readers of the original disclosure. Consistent with this, I find articles are much more

likely if the AAER alleges the company provided a material publicly misleading statement or

filing.

The press prefers stories that can be personalized and involve controversy, conflict and

deviant behavior. Frauds that involve management profiting due to insider trading, hidden

compensation or plain theft more easily lend themselves to such a spin. Consistent with this, I

find that frauds that are accompanied by such actions are more likely to result in an article.

As a final set of analyses, I examine the 76 articles that identify an accounting fraud to

determine how the press undertakes fraud investigations. Sixty-six percent of the articles discuss

at least one of the items of fraud in the subsequent AAER, suggesting reporters work with

relatively accurate information. The articles show that the press relies on a large range of

sources for information, the most common being analysts, followed closely by litigation/court

3
actions and auditor changes. Only a small number of the analysts cited worked in sell-side

positions, thus at the very least the press coverage is likely the first time this information is

broadly disseminated. While it is difficult to determine whether the reporter had already

identified the company and then found these sources, the fact that so many articles occur directly

following law suits and auditor changes suggests that reporters use these sources to identify firms

that are potentially committing accounting fraud. Regardless of the initial source, most articles

indicate the reporter collected additional information to complete the story. This suggests an

interesting informational interplay between several of the intermediaries and institutions that

govern society.

Turning to the type of publications, I show that these articles occur across many different

types of publications. While national business publications (such as the Wall Street Journal and

Business Week) publish the greatest number of the articles, many also appear in local media (LA

Times, Miami Review), electronic media (Dow Jones News Service, Bloomberg) and trade

publications (Boating Industry, Business Insurance). Only two of the articles appear in national

nonbusiness periodicals, indicating that acting as an accounting watchdog is likely not a priority

for these entities. Only four of the articles are featured on the front page or magazine cover,

while 15 more are on the front page of a back section of the paper. This provides further

evidence that the press may not view the accounting watchdog role as one of its primary

activities.

Local journalists may have an advantage in procuring information regarding a company,

however, only 24 of the 76 articles are either published in a co-located paper or have a local

byline (i.e. local stories). To further explore whether information sources are related to the

characteristic of the press/reporter, I interact the source of information cited with both the type of

4
publication and whether the story is local. Reporters from all types of publications rely on a

wide range of sources, though the national business press relies relatively more on analysts.

Further, articles published in a local paper (i.e. same area as company) or with a local by-line are

less likely to use analysts as a source and are more likely to be written without providing any

primary source for the information. If one assumes that articles without a source are more likely

to represent “pure” investigative reporting, this provides some evidence that the local press is

more involved in active investigations.

The main purpose of this study is to provide evidence on the role of the press in the capital

markets and society. The evidence indicates which type of firm the press is likely to effectively

monitor and how the press interacts with other information intermediaries and institutions that

help to facilitate the functioning of the capital markets. As Jensen (1979) points out, the press is

an integral component of our information environment and any complete theory of that

environment will require a thorough understanding of the role of the press. While the press

likely serves as important an information function as auditors and analysts, there have been

hundreds (perhaps thousands) of published studies on those information intermediaries and

relatively little on the press.

The evidence provided in this paper is also of interest due to the increasing use of the press as

a control variable in accounting and economic studies (Frost, Gordon and Hayes, 2001;

Bushman, Piotroski and Smith, 2003; Haw, Hu, Hwang and Wu 2003). With a greater

understanding of the press, we can more clearly determine whether these variables effectively

capture the constructs they are meant to measure.

5
 This paper proceeds as follows: section 2 discusses related literature and motivates the

research question, section 3 discusses the sample selection and collection of press coverage,

section 4 provides results and section 5 concludes.

2. Related Literature, Motivation and Research Question

2.1 The Process of Reporting

The press is a vital information intermediary in the economy (Owen, 2002). The

choice of items covered and investigated impacts the information set, actions and opinions of a

large portion of the population (Islam, 2002). Given this impact, understanding how the press

executes these tasks is critical in developing a comprehensive theory regarding how economies

and societies function (Jensen, 1979).

The process of reporting consists of several activities, including broadly disseminating

publicly available information, analyzing known information, creating “new” information

through investigative reporting and synthesizing all of this information into a framed story.

Differing stories require differing mixes of each activity. Covering a standard earnings

announcement may chiefly entail repeating the information disclosed by management with little

additional analysis. In this case, the press’ primary function is to more broadly disseminate

information. However, reporters still need to determine which information is newsworthy and

whether the original announcement suggests more detailed analyses is warranted. On the other

end of the spectrum, a reporter examining fraud may begin by examining public press releases

and financial statements, but is likely to supplement with less public sources, such as buy-side

analysts or others familiar with the company. For example, reporters working on an expose of

Enron used internet job sites to identify ex-Enron employees for interviews (Behr, 2002).

6
These activities are more investigative, thus the story will likely include information not

previously available to the general public and change the context of previously available

information.

In the process of reporting, the press interacts with the other information intermediaries in

society, such as analysts and auditors as well as with regulatory and punitive bodies like the SEC

and attorneys. The media often uses these entities as information sources while adding value by

making this information broadly available and synthesizing it with insights gained from other

sources to create “new” information. However, the media also serves as a source of information

for these entities. For example, short-sellers may take a position based on speculation of an

article being written and the SEC cites the media as providing leads for investigations (Kurtz,

2001, Feroz, Park and Pastena,1991).2

2.2 Related Literature

The existing academic research on the relation between the press and commerce is

relatively limited (Zingales, 2000). Of course, there are many studies that use the press to

examine the market response to significant events (for example, Niederhoffer, 1971 and Cutler,

Poterba and Summers, 1989). While the goal of these studies is to provide evidence regarding

the drivers of market movements, they also provide indirect evidence regarding the importance

of the press in covering an event.3 In a similar manner, studies of the market response to

fundamental analysis of companies by Abraham Briloff, an accounting professor at Baruch

2
Kurtz, (2001) cites an example of institutions taking positions in AOL based on a reporter asking questions that
suggested a negative story would be forthcoming. The concern that such activities occur is great enough that
publications such as the Wall Street Journal have enacted policies that forbid reporters from providing information
during interviews which may allow markets participants to determine the topic and/or tone of the article (Smith and
Emshwiller, 2003).
3
Both studies cited use coverage in the New York Times to identify news events, thus they jointly test the market’s
response to news and the New York Times’ ability to assess news. These studies find some evidence of a relation
between a news event and market responses. However the events are not able to explain the majority of variation in
price movements, suggesting some non-press covered information is moving the markets.

7
College, were undertaken primarily to investigate the relation between fundamental analysis and

market efficiency (Foster, 1979 and 1987; Desai and Jain, 2000). However, all of the Briloff

articles were published in a leading business magazine (Barron’s). Thus, studies of their impacts

also provide evidence regarding the press’s ability to impact equity valuations.4

Several studies provide more direct evidence on the relation of the press and business by

suggesting the press’ influences how other institutional entities fulfill their role. For example,

negative press coverage increases the likelihood that auditors will issue a going concern opinion

(Joe, 2003).5 Similarly, press coverage following the collapse of Drexel appears to have

exacerbated a political backlash against insurance companies heavily invested in junk bonds

(DeAngelo, DeAngelo and Gilson 1994, 1996). That backlash cumulated in California state

regulators seizing the companies despite liquidity ratios that were equivalent to those of similar

firms that were not invested in junk bonds.

Finally, there is a growing literature that uses cross-sectional variation in the national

characteristics of the press to investigate the role of the press in corporate governance,

governmental actions and development (Dyck and Zingales, 2002; Stromberg, 2002 and

Djankov, McLiesh, Nenova and Shleifer, 2002). This literature generally concludes that the

press can impact the political and economic make-up of a country.

Combined, the above studies provide important evidence that the press has a significant

impact on economic perceptions and actions. However, they do not examine how the press

decides which items to cover or how it collects information for analysis. Given the demonstrated

ability of the press to impact economic outcomes, understanding the factors that impact reporting

4
As Desai and Jain point out, the price drops are sustained over time, suggesting the analyses and publication
simply accelerated the market’s recognition of the underlying decline in value of the firms.
5
As discussed in Joe (2003), the cause of the increased chance of a going concern opinion is controversial and
appears to be at least partially due to an overreaction to redundant information on the part of the auditors.

8
is of obvious importance. Further, despite our lack of understanding of the reporting process,

measures related to the press are becoming increasingly common in accounting research (Frost,

Gordon and Hayes, 2001; Bushman, Piotroski and Smith, 2003; Haw, Hu, Hwang and Wu 2003,

Barton 2003)). Such measures could be formed and used more effectively once we achieve a

greater understanding regarding how the press functions.6

2.3 The Press as a Watchdog for Accounting Fraud

As previously mentioned, the press performs several functions including the broader

dissemination of public information, analysis of such information and investigative reporting. In

this paper, I focus on one aspect of press reporting – acting as a monitor or “watchdog” for the

public interest. This role is frequently referred to as one of the most important functions of the

press (Serrin and Serrin 2002; Islam 2002; Djankov, Mcliesh, Nenova and Shleifer 2002; Dyck

and Zingales 2002). Watchdog journalism occurs as “journalist ask penetrating questions at

every level of the public and private sectors” (Neiman Foundation project on Watchdog

Journalism web page) with the goal of “…holding to account society’s power brokers and rule

makers…” (Lewin 2002 pg. 19). This activity is viewed as quasi-governmental in nature and has

led to the press being referred to as the “fourth branch” or “fourth estate” of the government

(Groseclose and Milyo 2003). The watchdog process relies heavily on investigating and

analyzing non-public information. This “new” information is combined with publicly available

information in a way that highlights potential problems. Watchdog journalism in business

reporting is somewhat unique in that SEC filings and other publicly available information

provide a rich starting point for the process. However, as with other watchdog reporting, it still

6
As an analogy, the large amount of work performed in the analyst literature has increased our understanding of the
role of analysts in the institutional structures that make up the business world as well as greatly enhancing our ability

9
requires a reporter to identify the issue, improvise to collect other supporting information,

synthesize this information to make the issues in the story clear and then bring it to the attention

of the general public (Keller, 1998).

My study specifically examines the press’s role as a watchdog in the initial identification

and communication of accounting fraud and abuse. That is, I examine whether the press

identifies accounting irregularities prior to a public admission by the company or announcement

of an SEC investigation into the matter. This activity is a clear example of the press combining

information sources to highlight a potential abuse of a company’s duties and powers. As such, it

is perhaps the most direct example of the press acting as the early watchdog in business

reporting.7

Accounting fraud is an important event in evaluating companies, and thus also an

important news event. Further, it has been argued that the press chooses stories based on

entertainment value (Jensen, 1979). The often extreme actions, tensions and personalities

involved in accounting fraud combine to create a compelling story. For example, one popular

journalism text, Jamieson and Campbell (2001 pages 41-52) defines a newsworthy event as

having five characteristics: 1) can be personalized, 2) dramatic, violent and conflict filled, 3)

actual and concrete, 4) novel and deviant, and 5) linked to issues of ongoing concern.

Accounting fraud is one of the few business related stories that meets all of these criteria.

to use analyst related data in research regarding other issues.

7
As with many jargon terms, “watchdog reporting” has a broad and somewhat loose definition. All of the working
definitions seem to include the need for critical thought and question asking. Some also include the reporter being
the first broadly public source to address an issue. Others would expand watchdog reporting to the follow-up that
occurs after an issue is initially identified. Under those definitions, the in-depth coverage that occurs for many
companies following admission of a fraud would also be a form of watchdog journalism. Regardless of the
definition, follow-up coverage is an important part of the monitoring function played by the press. Examination of
that role is beyond the scope of this paper, but would be an interesting inquiry for future research.

10
Combining this with the press’ strong history of muckraking in general, it is reasonable to expect

the press will devote substantial efforts to identify and analyze accounting fraud.

However, there are countervailing pressures that may reduce the press’ activities as

watchdogs over firms. First, a large portion of the press is non-government owned and thus a

market participant themselves. As such, their interests are likely to be aligned with those of the

market, making it less likely that they desire to cause concern regarding this market or to be

scrutinized themselves (Herman, 2002).8 These apprehensions can be intensified by affiliations

with parent companies or advertisers who may be harmed by investigative reporting (Herman,

2002; Jamieson and Campbell, 2001). The press also may censor stories in an attempt to keep

ongoing sources available for future information (Jensen, 1979; Herman, 2002). Individual

reporters may also be concerned about retaining their job and/or future employability if

representatives of companies complain to their editors (McNair, 2002).9 Finally, the press also

has to be concerned about offending its readers, which are likely participants in the market

themselves (Jamieson and Campbell, 2001).10 Overall, these forces suggest that the press is

likely to face strong pressure opposed to uncovering accounting improprieties.11 Accordingly,

the press may choose not to act as a watchdog by providing initial or early information regarding

8
Jamieson and Campbell (2002 – pg 62-63) quote Peter Silverman, the business and financial editor of the
Washington Post: “Newspapers themselves are among the most secretive and the most protective about the facts
and figures of their own business. They are not likely to ask others to do what they are unwilling to do themselves.”
9
McNair presents several examples of attempts by businesses to silence reporters. While many were not successful,
in at least one case a reporter was fired shortly after company complaints about coverage (McNair, 2002 pg. 14).
10
As an example of a member of the press noting this constraint, Bill Wasik, an editor of Harpers, wrote an article
condemning business reporting as having devolved into nothing more than personal-finance journalism. He argues
the following :“That investigative reporting of Enron was carried out in inverse proportion to the company’s stock
price cannot be blamed merely on trend-focused journalists, or on bought-off editors or publishers. Rather, it is
intrinsic to the culture of personal-finance journalism. When a company’s fortunes seem poised to enlarge
indefinitely, the interests of all potential sources- the company, the analysts, large investors-are aligned not only
with one another but with the interests of the reader, who is assumed to be a shareholder or potential shareholder.”
(Wasik, 2003) (Italics in Original).
11
It is interesting to note the similarity between these forces and those studied in the analyst and auditing literatures.

11
the problems, but rather may wait until a fraud has been uncovered and then fulfill the need for

interesting business stories by writing ex-post articles regarding the frauds.

2.4 Research Question

The first issue addressed in this paper is the degree to which the press fulfills watchdog

role. I define this role as the press identifying (or “catching”) an accounting violation prior to a

public announcement by the firm or SEC regarding an examination or restatement for the firm.

As I discuss more fully in the data section, I include firms in my sample based on being found

guilty of accounting violations or fraud by the SEC. Thus, examining the proportion of these

firms identified by members of the press prior to the official announcement of an accounting

issue allows me to provide evidence regarding the press’ ability to identify and its willingness to

publish suspected fraud.

Next, I examine the characteristics that lead to the press identifying and publishing an

incident of accounting fraud in some instances while missing or staying silent in others. I expect

reporters to maximize the benefits to the publishing agency. First, the press would prefer articles

that will be of interest to a large group of readers (Stromberg, 2002). Those articles will attract a

large group of readers, resulting in higher subscription revenues and a larger reader base to offer

potential advertisers. Since subscriptions and advertising are the two primary revenue generators

for the press, it is obvious that the choice of what to publish is economically important. Further,

the press will want stories readers find memorable, thus increasing the chance the readers will

use the publication in the future (Mullainathan and Shleifer, 2002). Additionally, as with any

business, the press must also consider the cost of undertaking an activity. Firms with a richer

information environment can be analyzed more easily, and thus will be less costly to cover.

12
Further, certain types of accounting violations may be easier to detect, and thus be more likely to

be covered.

As is often the case, it is difficult to separate the impact of cost and benefits in much of

my empirical analyses. Accordingly, I discuss the empirical implementation of the predictions

provided by costs and benefits together.

I examine whether firm characteristics impact the likelihood of fraud being caught by the

press. I expect that large firms are more likely to have a large number of stakeholders interested

in the firm’s activities, and thus present an opportunity to attract a large reader base. Further,

these firms are more likely to have a rich information environment due to both stakeholder

demand for information and the number of other information intermediaries following the firm.

Thus, the benefits of coverage are higher and the cost of analyzing these firms is lower.

Combined, this suggests that large, heavily followed firms are more likely to have any

accounting fraud identified by the press. I proxy for these attributes by using market value,

analyst following and overall press coverage (details of these items are discussed more fully in

the sample and results sections).

A firm’s advertising spending is another characteristic that may impact the likelihood that

a critical article is written. Given the importance of advertising in the overall revenue generation

of the press, publishers may be hesitant to print articles that will offend large current or

prospective advertisers (Reuter and Zitzwitz, 2003). Articles alleging inappropriate accounting

have a high potential for upsetting any affected advertisers. Thus, I predict that firms that have

the potential to be large advertisers are less likely to have unfavorable articles written about

them. I proxy for advertising status using an indicator variable that is coded as 1 if the firm is an

13
industry that was in the top 15 advertising industries (according to Advertising Age) in each of

1985, 1990, 1995 and 2000.

Next, I expect that the type of fraud impacts both its appeal to individuals (based on the

five items that indicate newsworthiness discussed above) and the ease with which the press can

detect it. Accordingly, I examine whether fraud characteristics impact the likelihood it is caught.

I expect that fraud that is extensive in scope, based on items manipulated or time involved, is of

greater interest to the public and easier to detect. Thus, I investigate whether the number of

items of fraud indicated in the AAER is related to the likelihood of being identified by the press.

Similarly, I examine whether fraud that occurs over a longer period is more likely to be

identified. I also expect the press to be more likely to identify a fraud that involved a materially

misleading public statement (such as an earnings announcement or discussion of new contracts).

By their very nature, such statements are likely to attract scrutiny and communicate items the

company expects will catch the public’s attention, suggesting broad appeal to consumers of the

press. Many accounting frauds occur concurrent with management misappropriation of funds,

either through illegal insider trading, manipulation of bonus plans or just pure theft. While it is

unclear whether such items will be any easier for the press to detect, they clearly increase the

degree to which the article can be personalized, is dramatic, has conflict, and covers deviant

behavior. As such, the press is likely to consider such stories more compelling. Finally,

although I have no ex-ante expectations, I also investigate whether the two most common types

of fraud (revenue manipulation and asset overstatement) are related to the chances of being

identified.

My final analyses examine the content of the articles alleging accounting concerns. The

goal is to provide an understanding of the information sources the press uses, which type of press

14
identifies fraud and whether those two items are related. Of course, isolation of sources based on

the ex-post article is difficult as the press may initially receive information from one entity which

does not want to be identified and so in the process of writing the article a subsequent source

(which is willing to be cited) may be used.

3.Research Design and Sample

3.1 Sample Creation and Accounting and Auditing Enforcement Releases

To investigate what type of firm and fraud lead to the press identifying an accounting

fraud, I must identify firms that have committed accounting fraud. Consistent with prior work, I

use the SEC Accounting and Auditing Enforcement Releases (henceforth AAER) to build a

sample (Dechow, Sloan and Sweeney, 1996; Bonner, Palmrose and Young, 1998; Beneish, 1999

A, 1999 B). There are differing opinions regarding why AAER are issued. While some believe

they are to address current trends observed by the SEC (Feroz, Park and Pastena, 1991), others

believe they are issued for high profile issues that will enhance the stature of the SEC (DeFond

and Smith, 1991). In any case, there is wide agreement in the literature that AAERs generally

represent egregious violations of the SEC standards. As such, they provide the basic criteria

needed to begin developing my sample: an observable sample of firms that have committed

accounting fraud. AAERs have the added advantage that they provide other detailed

information, such as a discussion of the violations, summary of findings and a time line of the

violation. This in-depth characterization is helpful in developing variables to investigate how the

type of fraud impacts press coverage.

Use of the AAERs also has disadvantages. First, AAERs likely represent extreme

violations. If these firms have characteristics that vary systematically from firms that commit

fraud but do not receive an AAER, then the result may not generalize into the population in

15
general. Second, Feroz, Park and Pastena (1991) report that an SEC official indicated that

approximately 1/3 of its leads are a result of scanning the financial press. It is not clear whether

that entails only articles that allege fraud, or simply that the press is one source for collecting

information. However, it is possible that a portion of these firms would not have been identified

without the press article. Assuming that all such articles are read by the SEC, the findings in this

paper would likely overstate the proportion of frauds in the general population that are identified

by the press.

AAERs are identified using the SEC web site.12 AAERs are issued for a wide range of

violations. While accounting fraud is a primary reason for an AAER, they also frequently

include illegal insider trading, violations of SEC requirements by auditing firms and other illegal

acts. For the purposes of this study, only AAER that include a substantial accounting fraud are

retained.13 Subsequent analyses rely heavily on data from the Factiva news service, such as

identifying the date a firm issues a press release that it is under investigation or restating

earnings. The two primary services for firm press releases are PR Newswire, with coverage

beginning January 2, 1985 and Business Wire, with coverage beginning July 28, 1998.

Accordingly, AAERs with violation periods that began prior to January 2, 1985 are excluded.14

12
At the time the search was performed, the SEC web site did not include a single comprehensive listing of AAERs.
Thus, the site’s search function was used. This search function does not adjust for items such as hard carriage
returns and wrapping lines, so it is necessary to perform several searches using variations of the words Accounting,
Auditing, Enforcement and Release. While the text of most AAER are available on the SEC website, in several
instances it was necessary to use LexisNexis to obtain a full text version of the AAER. As a robustness check, the
AAER identified were compared to those in the October of 2002 GAO document “Financial Statement
Restatements, Trends, Market Impacts, Regulatory Responses , and Remaining Challenges.” There were no
discrepancies noted. However, it is possible that the sample used in this paper missed some issued AAER’s. I do
not anticipate this having a systematic impact on the results.
13
For example, I noted several AAER that involved payments that violated the foreign corrupt practices act. These
AAER also noted accounting violations related to how the payments were recorded. However, the magnitude of
these payments was always minor in relation to the statements as a whole.
14
Robustness tests indicate the results are similar if AAER with violation periods prior to July 28, 1998 are
excluded.

16
As shown in Table 1, this results in a sample of 263 AAERs with violation periods that end over

the years from 1987 to 2002.

Information from each AAER is coded including the firm name, time period of the

violations as well as number and type of violations alleged by the SEC. Table 2 provides

summary descriptive statistics of those data. Panel A indicates there is a reasonable cross-

sectional variation in the length of violation periods. While 30% of the violations last one year

or less, 5 violations lasted seven years or longer.15 Consistent with prior research on AAERs,

Panel C shows that revenue manipulation is the most common form of violation (49% of the

sample) (Feroz, Park and Pastena, 1991). Asset overstatement is the second most common

violation (35% of the sample). Other common violations include providing materially

misleading public statements/filings (26%) and manipulation of reserves (18%). These

percentages sum to greater than 100% because the SEC frequently identifies more than one

violation per firm, thus the categories are not mutually exclusive. In untabulated analyses, I find

that 12% of the firms do not any of the tabulated violations. The untabulated violations consist

of various items. For example, delayed recognition of losses occurred in 8% of the sample (and

reduces the non-explained group to 6%). Panel B also shows that 24% of the firms have a

violation that indicates management profited from the fraud through an action such as insider

trading, gaming of compensation contracts or outright theft.

Panel C shows that the number of violations noted in the AAERs vary from one (16% of

the sample) to seven or greater (1% of the sample). The number of violations reported on each

AAER did not appear to follow standardized formatting, thus this variable likely includes a high

degree of noise.

15
Untabulated analyses find that 18 of the firms have violations for exactly one year, while only 8 firms have
violations that span 1 quarter or less.

17
 As a robustness check, all variables were interacted with the year of the violation to

assure there are no trends over time that may impact the use of these variables in later analyses.

No such trends were noted.

3.2 Press Coverage

Having identified firms with an accounting fraud, the next step is to collect press

coverage of those firms. Coverage is collected over several time periods with searches designed

to create different variables. Figure 1 provides a timeline of the various searches. All press

searches are performed on Factiva, which is Dow Jones’ replacement for the Dow Jones News

Retrieval service. Factiva covers nearly 8,000 sources of information including all of the major

wire services (for example, PR Newswire, Business Wire, Dow Jones, Rueters, AP), major US

business publications (Wall Street Journal, Barron’s, Forbes, Fortune, Business Week), national

and regional newspapers (New York Times, Washington Post, Los Angeles Times, St.

Petersburg Times) and trade publications (Computergram, Boating Industry).16

The first searches examine press coverage over the period that the accounting fraud

occurred (i.e., the infraction period). This information is used to assess the intensity with which

the press examined the firm. The searches are performed using multiple variations of the firm

name within one search, allowing for both the full corporate name and diminutives. Several

firms changed names during the sample period. Their searches were adjusted accordingly.

Because the goal is to identify press coverage, the wire services that directly reprint managerial

news releases (PR Newswire and Business Wire) are excluded from this search. Similarly, the

summaries that simply pick-up a list of firm names and an item such as a large traded volume

16
As a robustness check, several firms were randomly selected and press searches were run on Factiva, Dow Jones
and LexisNexis. The results found Factiva to be a superset of Dow Jones. While a few items appeared on
LexisNexis that were not on Factiva, they were not from major publications and did not include information
excluded by sources on Factiva. Further, Factiva generally had a broader set of articles than did LexisNexis.

18
during the day are excluded.17 Searches cover the entire text of all documents. The resulting

number of press articles was then deflated by the length of the infraction period (in months) to

standardize press coverage across firms.18 As shown in Table 3, the distribution of the number

of articles per month varies greatly. While 42 of the firms average one or fewer articles a month,

7% are included in more than 100 articles per month during the infraction period. However,

most firms have between 2 and 20 articles per month.

Many of these articles likely cover the same event. For example, one earnings

announcement may lead to multiple articles within a few days. Since the goal was to identify

overall press coverage, these duplicates were retained. To provide some sense for the underlying

amount of information, a similar search was performed over the same time period including only

the management press wires (PR Newswire and Business Wire). Some companies are named in

press releases of other companies (for example, Microsoft is often mentioned as a client or

business partner). Thus, this search requires the company to be named in either the headline or

the first paragraph. Not surprisingly, the managerial press release intensity is highly correlated

with the press intensity, with Pearson(Spearman) correlations of .83 (.73) (untabulated).

However, as Table 3 shows, when compared to the press intensity, the managerial press release

intensity tends to be more tightly clustered with many fewer releases. In fact, 42% of the firms

average one or fewer a month and 91% of the firms have fewer than 5 press releases per month.19

Finally, the search for articles that “caught” the accounting fraud is performed. That

search was run from the beginning of the infraction period until the date of the AAER being

17
These summaries include the DJ Highlights, News Highlights, V-alert, P-Alert, California Summary, Southeast
Summary, Southwest Summary, Dow Jones Corporate Economic News Summary, Recap of Dow Jones Special
Reports, Wall Street Journal Earnings Summary and the International Calendar of Corporate Events. This list was
compiled based on performing unrestricted searches and identifying these types of items as they appear.
18
All reported results are similar if the undeflated variable is used.
19
The goal of the paper is to examine press actions, thus press intensity is used throughout the remaining paper.
However, results are similar when managerial press release intensity is substituted.

19
issued. However, articles were excluded if written after a public announcement by the firm of

either an SEC investigation or accounting restatement. There are often several years elapsed

between the infraction period and issuance of the AAER. Accordingly, it was necessary to

reduce the number of articles somewhat by using a search string. This search string was

designed to be comprehensive in identifying articles that question the firm or its accounting

while reducing the number of spurious articles to be read.20 Even with the search string, many

companies had hundreds of articles to be read, thus a relatively comprehensive set of articles was

examined for each firm. If a caught article was identified, only the first article was retained.

Data on that article was collected including the publication, author, type of article and cited

source of information. The decision of whether an article “caught” an accounting failure is

obviously judgmental.21 An research assistant who was not aware of any of the hypothesized

relations performed all coding to avoid any coding bias.

4. Results

4.1 Characteristics of Firms “Caught” by the Press

The first analysis examines the proportion of firms for which the press identified the

questionable accounting prior to a public announcement of an accounting problem. As shown in

20
The search included both the formal and shortened versions of the company name appearing in the same
paragraph as any variation of: accounting, audit, fraud, illegal, illicit, insider trading, investigate, overstate,
understate, probe, quit, quits, quitting, resign, restate, revenue, revenue recognition, rumor, law and suit within two
words of each other, opinion and withdraw within three words, short and sell or stock within three words. It also
included the company name and any words from the following group: adjustment, compensation, doubt, dubious,
financial, number, officer, recognition, record, reserve, Securities and Exchange Commission and SEC combined
with any words from this group: corrupt, conceal, credible, debacle, deteriorate, difficult, discrepancy, dishonest,
failure, false, falsify, fear, improper, inconsistent, ills, irregular, misappropriate, mislead, misrepresent, negative,
offense, question, sell-off, shortfalls, skeptic, suspicious, trouble, unexpected, unsupported, violated, weak, woe,
worried, write-off. Several companies with no caught article identified by the primary search were searched without
the limiting string. No additional caught articles were found.
21
In fact, even the press cannot agree on whether a specific article effectively identified misdeeds at a firm. For
example, in the summer 2002 Neiman Reports there are three articles regarding press coverage of Enron. Madrik
contends that reporters did not look skeptically at Enron, and in fact helped to perpetuate many of its practices. On
the other hand, Behr argues that negative coverage was minimal, but points to a March 2001 Fortune article. Steiger

20
Table 4, 76 of the 263 firms (approximately 29%) were identified prior to the public

announcement. Table 4 interacts the percentage of firms caught with the last year of the

violation. The percentage varies over the years with the highest percentages in 1993-95 and

2001-2002. Both of these periods occur during recessions that happened after periods of strong

market performances and thus may suggest the economy has an impact on the likelihood of an

article. However, given the relatively small number of years examined this observation should

be viewed with caution. Overall, the increases and decreases in caught articles during the sample

period suggests that, with the possible exception of responding to macroeconomic cycles, the

press’ actions as a watchdog of accounting fraud have not changed over the time of my study.

Having established variation in whether the press identifies accounting fraud, I next turn

to investigating cross-sectional factors associated with whether a given firm is identified. As

discussed in section 2.3, I expect that identification will vary both with characteristics of the

firm’s information environment and with the type of fraud committed.

I expect that firms that have a rich information environment are more likely to be

identified due to the fact that this rich environment reflects a high degree of interest in the firm in

general (thus increasing the benefits to the reporter/publisher of an article) and lowers the cost of

investigating the firm. I measure the information characteristics with three variables. First, press

intensity (PRESSINT) captures the amount of attention received in general by the press. Recall

that PRESSINT is calculated by identifying all articles written over the period of the violation

and then deflating by the length of the violation in months. The second information variable is

analyst following (ANALYST) measured in the last month of the violation period. Analyst

following is highly skewed. Consistent with prior research, I transform this variable into a

(a managing editor at The Wall Street Journal) argues that Enron’s misdeeds were uncovered in October of 2001 by

21
dichotomous variable coded as 1 if the firm has analyst following in the top quartile of the

distribution of my sample (Bradshaw, Bushee and Miller, 2003).22 The third variable is the

market value (MV) of the firm. Consistent with prior research (Feroz, Park and Pastena, 1991), I

find that many firms cannot be identified on CRSP (82 or 31% of the sample, almost evenly

distributed between the caught and not caught subsamples). Obviously, missing firms are

excluded in the MV analyses. However, they are included in all other analyses. As might be

expected, untabulated analyses finds that the three variables meant to capture information

environment are highly correlated. In fact, the lowest correlation occurs between ANALYST

and PRESSINT with Pearson (Spearman) correlations of .64 (.72). Further, the Cronbach’s

alpha for the three variables is .83, which exceeds Nunnally's (1978) suggested value of 0.70 for

a reliable index. Finally, factor analysis of the three variables finds only one factor. When

multiple factors are forced, the first factor loads with an eigenvalue of 1.72; all forced factors

have negative eigenvalues, implying that they reduce the ability to explain variation in

comovement. Given this high correlation, I do not combine these variables in my logit

regressions. Rather, I run the logit three times, once with each information environment

variable.23

The press may hesitate to write articles that will upset large advertisers for fear that this

will impact future advertising from those firms. I examine whether the press less often identifies

“relentless, careful, intelligent work of two Wall Street Journal reporters.”

22
This variable is calculated using IBES data. If a firm could not be identified on IBES, analyst following is
assumed to be zero. Results are similar if such firms are discarded. Results are also generally similar, although the
variable becomes statistical insignificant in some tests, if the variable is left in its continuous form. As a final
robustness check, I also include all observations but use logs of analyst following. Again, results are similar.
23
If the factor square is used in place of the information variables in the logit, the coefficient on the score is positive
and significant. The findings on all other variables remain the same as those presented in the tables. I do not use the
factor score in the primary analysis due to the large number of lost observations caused by missing market values
and the subsequent differences in findings for MV vs. those of PRESSINT and ANALYST.

22
firms in a high advertising industry as having questionable accounting by including an indicator

variable if the press is in a high advertising industry.

Some aspects of the accounting fraud may influence whether the press detects the fraud.

First, I expect more egregious frauds to be more likely to be detected. I measure the magnitude

of the fraud using two items: the number of violations noted (NUMVIOL) and the length, in

months (LENGTH).24 Second, I expect frauds that involve material public misrepresentations

and misleading statements (through a press release or financial statements) (MISLEAD) to be

more likely to attract press attention since they are highly visible public statements.25 Third, I

expect that fraud that involves some form of management misappropriation of funds will be

viewed as more intriguing stories by the press. Accordingly, I expect that reporters and

publishers will more likely look for and follow-up on stories regarding such items. A company

is coded as having such misappropriation (STEAL) if the AAER included censure for insider

trading, illicit payments, undisclosed compensation or pure theft. Finally, as revenue recognition

(REV) and asset overstatement (ASSETOVER) are the two most common forms of

manipulation, I also include indicator variables for each in the analyses. I do not predict a sign

for these variables.

Table 5 Panel A provides univariate and Panel B provides logit analyses of the difference

in these variables between firms caught vs. not caught by the press. Overall, the results indicate

that information environment variables are related to the likelihood that the press will publish an

article regarding accounting fraud. Caught firms have a significantly higher mean and median

level of PRESSINT and ANALYST. Both variables are also significant in the logit analyses that

24
Ideally, I would also include a variable capturing the dollar value involved in the fraud. Unfortunately, the AAER
are not always clear on magnitude. Further, the types of fraud vary – making aggregation somewhat problematic.
25
For example, several of the AAER firms’ frauds included making public statements that large contracts had been
signed. These same firms often then manipulated revenue consistent (and/or concurrent) with these claims.

23
control for other items. However, the MV variable is significantly greater for caught firms at

conventional levels in the median, but the mean and the logit results slightly exceed the 10%

cutoff. To investigate whether this result is due to the smaller sample size for MV, I reperformed

all tests for PRESSINT and ANALYST restricting the sample to firms with MV data. The

results for both variables continue to hold at close to identical p-values, indicating it is not

sample size alone that creates the weaker results for MV.

While the above results suggest firms characteristics have an important impact on

coverage, both the univariate and logit result fail to provide evidence that the press is less likely

to cover firms which represent the possibility of substantial current or future advertising revenue.

This may be due to noise in my measure due to the use of industry level data and /or the fact that

many high-advertising industries may not advertise in the business press (for example, consumer

products companies may not view the readers of the business press as their target market).

However, findings from concurrent research suggest the cause may be more complex. Reuter and

Zitzwetz’s (2003) study whether advertising spending by mutual fund companies impacts the

likelihood that the fund will be recommended in a financial publication. They find no evidence

of bias for The Wall Street Journal or New York Times, but evidence consistent with bias for

Kiplinger’s, Money and Smartmoney. Interestingly, many of the caught articles in my sample are

found in The Wall Street Journal and its related Dow Jones Newswire (14 articles each), and The

New York Times also provides one article. There are no articles in the three magazines with

evidence consistent with bias in the Reuter and Zitzwetz study. This suggests that while

24
advertising spending may impact some publications, the remaining outlets continue to serve a

monitoring function.26

Results regarding the type of fraud impacting whether an article is written are mixed.

Frauds that involve public misleading statements are more likely to result in an article being

written. Frauds that involve misappropriation of funds are also more likely to result in an article.

However, there is little support for the severity of the violation impacting the likelihood of an

article being published. The length of fraud is insignificant in all analyses and numbered of

violations is significant only in the mean. While asset overstatement is significantly negatively

related to the chance of an article being written in the mean and median, it is insignificant in all

of the logit analyses. Revenue recognition is insignificant in all analyses.27

Combined, these results indicate that firm characteristics and the manner of fraud both

impact the effectiveness of the press as a watchdog over accounting fraud. To develop a greater

understanding of how the press goes about this watchdog role, I next use the articles on the 76

firms that were caught by the press to provide a descriptive analysis of the information used by

the press and the type of press performing this watchdog function.

4.2 Content Analyses of Sources Used and Type of Publications

Untabulated analyses finds that 50 of the 76 articles, or 66%, included information

regarding at least some of the specific items included in the eventual AAER. This high degree of

accuracy suggests that the press works with relatively informed resources.

26
This conjecture is also consistent with a Neiman Foundation survey of members of the press. The journalist
surveyed voted the New York Times and Washington Post as doing the best job of Watchdog journalism, followed
by The Wall Street Journal. The votes for other outlets were not given.
27
Correlation and factor analyses indicate the variables on the type of fraud are not highly correlated. However, to
assure the lack of results in the logits are not due to collinearity issues, I reperformed the logits including only one
AAER variable in each logit. NUMVIOL (REV) was borderline significant when ANALYST (PRESSINT) was
included as the control variable. All other results remained as reported.

25
 Table 6 examines the sources used by the press. In 18% of the articles no external

sources were cited. Some of these articles provide enough information to conclude they are the

result of press-initiated investigative reporting, but many of the others simply point out problems

with the accounting without providing the reason for investigating the firm.28 The remaining 62

articles include references to sources used in researching and writing the article. While being

cited as a primary source does not mean that it provided the information that initiated the article,

it does indicate that the party provided information or expressed an opinion that was useful to the

press in understanding and framing the issue. An understanding of sources used by the press

provides insight into the process of investigating and reporting a story.

Analysts are cited in 23 (30%) of the articles, making them the most commonly cited

source of information. In most cases it is difficult to determine whether such articles are the

result of press-initiated or analyst-instigated coverage. Limited information in the articles

suggests both occur. A further understanding of the relation between these two information

intermediaries is complicated by the fact that many of their activities are so highly related (both

collect information from a wide range of sources to provide information regarding the firm to

stakeholders outside the firm).

To investigate the relation of the press and analysts’ information, I examine the type of

analysts cited (non-tabulated). Since sell-side analysts’ role is to produce information that is

available to the public, I would expect that the press adds less value when including information

from these analysts (although the press could still add value by combining the public information

of the sell side analyst with information from other sources).

28
As an example of articles that indicate press-initiated investigative reporting occurred, a reporter noticed an ad to
sell specialized used computer equipment as part of liquidating a line of business. No company was identified, but
there was a local (Silicon valley) number. The reporter called the number and asked what company he had reached.

26
 My findings indicate that at the least the media is the entity that first alerts the general

public to the concerns. Seven of the analysts could be classified as sell side and therefore likely

to provide the information directly to the public. Only three of those articles refer to publicly

available analysts report. The remaining four refer to sell side analysts (two equity and two

debt), but with no mention of a written report regarding the issue. Conversely, the remaining 16

articles cite analysts affiliated with private information bodies. Five of those articles cite private

newsletters (three are Howard Schilit) that are normally reserved for wealthy individuals or

institutions. The remaining eleven articles identify analysts as either buy-side or short sellers (in

some cases no name is given, rather the author just refers to “shorts”).29 In three of these cases

the analysts have written letters to the SEC regarding their concerns and in one the analyst made

a similar statement on Dan Dorfman’s TV show, suggesting there may have been some public

information prior to or concurrent with the article. The other seven articles make no reference to

any other previous public pronouncements by these analysts.

The second most frequently cited source is legal cases, which occur in 15 (20%) of the

articles. They range from lawsuits on behalf of shareholders alleging malfeasance, to civil cases

involving wrongful discharge and even criminal cases investigating theft. Obviously, it is easier

to determine causality in this situation. The case is filed, leading the reporter to investigate the

company. While some articles appear to rely entirely on information in the case, most articles

He investigated the company and found they had made several recent public statements regarding the strong
performance of that sector of their business with no mention of liquidation.
29
This suggests the interesting question of whether buy side analysts/shorts use the media in an attempt to move
markets in their favor. The question is complex as it sometimes is even difficult to identify the line between the
media and these investors. For example, Cramer founded and is one of the major shareholders of The Street.com,
has been an active writer in several media outlets for years and is the managing partner of a large hedge fund.
Similarly, business-focused programs such as Squawk Box have a rotating position filled by different active analysts
or investors at all times (Kurtz, 2003). The SEC seems to have responded to this potential conflict by including in
Sarbanes-Oxley sec 501 a requirement that analysts disclose their current and recent past holdings in any company
discussed in a public forum.

27
include additional investigative reporting such as speaking to officers or customers of the

company. Thus, while the legal system may be the first to identify the situation, the reporter

usually adds further investigative services. Given the large number of cases filed in America

each day, reporters may also be fulfilling an important information screening function for the

public.

The third most common source is auditor resignations, cited in 12 (16%) of the articles.

Similar to the articles driven by legal cases, the reporters use the resignation as a cue to look

more carefully at the company and then develop information for their articles that goes beyond

that found in the 8-K auditor filings with the SEC.

In aggregate, these three sources (analysts, legal cases and auditors) provide information

for 66% of the total articles written. Given their role in capital markets, this provides a

beginning in documenting the relation among the press and other institutions called for in Jensen

(1979).

Although the remaining sources occur much less frequently, they still help to provide a

more complete picture of how the press investigates fraud. Four of the articles specifically refer

to non-standard filings with the SEC (for items such as share issuances and sales) as providing

the basis for writing the article. Three articles cite a customer as the primary source of

information and two cite industry insiders. Similar to the analysts, it is difficult to determine

whether the reporter approached these sources as part of an ongoing investigation or whether the

customer/industry insider approached the reporter. Finally, two articles cite an anonymous tip

and one information on a financial web site.30

30
An interaction of the level of specificity with the type of news finds that the source of information does not appear
to be related to the level of specification, at least for the three primary sources of information. While 66% of the
articles include specific reference to some portion of the fraud, 61% of the articles with analysts, 60% of those with

28
 Untabulated analyses find that the articles appeared in 40 different publications from 54

different authors. The Wall Street Journal and Dow Jones News Service both have the greatest

number of publications with 14 each, followed by the San Francisco Chronicle with 6 articles. 31

The remaining publications with more than one article are Business Week and Barron’s (each

with 4), Forbes (3), and Tulsa World, The Orange County Register, Computergram International

and the Miami Review (each with 2). Only one author has more than one article (Herb

Greenberg – who accounts for 5 of the San Francisco Chronicle articles); 18 of the articles do

not have an author in the byline.

As Jamieson and Campbell (2001) point out, any classification of type of publication is

subjective. With this caveat in mind, Table 7 provides more aggregated information regarding

the types of publications and articles that address accounting fraud. Twenty-nine of the articles

were published in national business publications such as The Wall Street Journal, Business

Week, etc. Twenty-three are from “local market” publications (for example, Chicago Sun-Times,

Miami Herald etc.) Fifteen of the articles are from electronic media (Dow Jones News Service

and Bloomberg). Seven are from trade publications (for example, Boating Industry, Business

Insurance). Finally, two are from nonbusiness national publications (USA Today and the New

York Times).32 While this descriptive data shows the importance of business publications in

serving as watchdogs for accounting fraud, it also shows that many other sources uncover and

publish articles regarding accounting fraud.

litigation and 67% of those with auditors include specific information. These values are economically and
statistically equivalent (p-value .93).
31
These two sources are related. Dow Jones News Service has over 20 regional offices that prepare articles that go
out on the Dow Jones Wire Service. These articles are also submitted to the editors of the Wall Street Journal who
determine whether some edited version of the article should appear in WSJ (Thompson, Olsen and Dietrich, 1987).
32
Factiva’s only provides limited coverage of the The New York Times during some portions of my sample time
period. Accordingly, the number of articles attributed to The New York Times may be understated. As a robustness

29
 While one would expect local journalists to have an advantage in procuring information

regarding fraud, untabulated analyses find that only 24 of the articles have a byline or are

published in a outlet that suggest a regionally based reporter. However, this may be understated

as a large number of the articles do not contain a by-line.

Forty-nine (65%) of the articles focus only on the company identified in the AAER,

suggesting that fraud articles tend to isolate companies. However, untabluated results find that

only four of the articles make the front page or magazine cover. An additional 15 articles are on

the front page of a back section and, of course, the 15 electronic articles do not have a location in

the paper. The remaining 42 articles are located somewhere in the middle of the periodical.

Overall, the article locations suggest that they are not considered primary news.

These descriptive data have shown that reporters rely on many sources for information,

but that analysts, lawsuits and auditor changes are the primary sources. It also shows that the

business press is the predominate outlet for “watchdog” articles, but that a significant minority of

the articles are published in other sources. In the final analyses, I examine whether any relation

exists between the sources and types of publications. To reduce dimensionality in these

analyses, I report the three primary sources of information (analysts, litigation and auditor

changes), an “other source” category and no source mentioned. I also reduce the categories of

publications by including the two national publications in the local market category.

As shown in Table 8 Panel A, reporters from each publication type rely on a wide-variety

of sources to identify accounting fraud. However, a few notable items appear. First, the national

business publications have a relatively high reliance on analysts. Second, electronic and trade

check, a research assistant used LEXIS/NEXIS to search 10 companies for which no article was found and 10
companies with an article from a publication other than The New York Times. No other articles were found.

30
publications always provide a source. While the Chi-square (not tabulated) is .06, the low

number of expected observations for each cell reduces the validity of this test.

Panel B shows the relation between a regional by-line or paper and the sources used to

examine whether proximity to the firm impacts the way the reporter investigates the firm.

Relatively speaking, locally based reporters write more articles without providing a source and

are less likely to use analysts. One interpretation of this is that locally based reporters are more

likely to undertake their own investigation of a firm while other reporters rely on information

intermediaries. The untabulated Chi-square is .11 and should be viewed with the same caveat

that the data is not very conducive to a Chi-square analysis. However, analyses by cell indicates

that the no source and analyst categories drive the near significant result.

4.3 Market reaction to articles

Finally, I also provide descriptive evidence regarding the market reaction to these

articles. This event-study faces several threats to its validity. First, the day on which the

information became available for many of the articles is difficult to determine. This threat is

relatively limited for wire services (which report real time) and newspapers (which may come

out after trading for the current day, but can be accommodated with the standard three day

methodology). Magazines, however, are frequently published in advance of the “issue date”,

thus introducing substantial noise. Further, in at least one instance management issued a

preemptive press release challenging an article more than a week before the article was published

(Bausch and Lomb with Business Week). Second, as shown in the Table 6, several of these

articles cite relatively recent events, such as lawsuits and auditor changes, that are likely to have

impacted returns as well. If these items occurred within the 3-day period used for most event

studies, then my results may be biased. Third, analysts are a commonly cited source of

31
information. They may provide concurrent advice to clients triggering the stock reaction.

Finally, my sample is small and was not designed for powerful tests of market response to

information. Given these threats, I present the returns results only as descriptive evidence and do

not tabulate them.

I was able to identify returns data for 56 of the articles. The average (median) three day

market adjusted reaction centered on these article dates is –9.1 (-5.5) %. Further, 75% of the

observations are negative, suggesting a pervasive view of the article release days as bad news.

Obviously, this is consistent with articles alleging accounting abuse causing a negative market

reaction. The same measure for the single day that the article was written is –6.5 (-2.9)%. All

of these amounts are statistically significant. Overall, this evidence indicates that these articles

provide new information to the market, but this evidence should be viewed with caution as

discussed above

5. Conclusion

This paper begins to answer Jensen’s (1979) call for a greater understanding of the press

in the markets, and of the “muckraking” function of the press in general, by examining the role

of the press as watchdogs for accounting fraud. It contributes to our understanding of how often

the press fulfills this role, which types of firms the press is more likely to effectively monitor and

how the press goes about the process of investigating accounting fraud.

Using a sample of firms sanctioned by the SEC for accounting fraud, I find that the press

is more likely to act as an accounting watchdog for firms with heavy analyst following and a

high degree of overall press coverage. I also find some support for the press being more likely to

publish articles on larger firms. Overall, this evidence is consistent with the press focusing on

companies that have a large public following and rich information set to draw on. Such a choice

32
is consistent with the press maximizing the benefits of investigative reporting by appealing to a

large reader base while minimizing the cost of investigating firms.

Consistent with this cost/benefit model, I also find the press is more likely to write an

article when the fraud involves a misleading public statement. Such statements may attract the

press’ attention initially, thus reducing search costs. They also have been viewed by many

members of the public and are obviously considered important by management. Thus, any

subsequent article should have a large prospective reader base. As further evidence of the press’

desire to write articles the public finds compelling, the press is more likely to write an article if

managerial misappropriation of funds has occurred. Such instances allow for a personalized and

conflict filled article, which are frequently cited as goals of a good article (Jamieson and

Campbell, 2001).

Inconsistent with my prediction that more egregious violations are of more interest to the

press, I do not find that the extent or the length of the violation are related to the occurrence of an

article.

Next, I provide descriptive data about the sources used by reporters and the type of press

that undertakes the watchdog role. I find that the press relies heavily on information from

analysts, auditors and lawsuits. This suggests an ongoing interplay between capital market

information intermediaries and institutional players in society. While I find that the national

business press takes the lead in performing this watchdog role, a significant portion of the

articles appear in regional and trade publications as well as on electronic media. Finally, most of

the articles do not suggest any local presence of the reporter. However, the few locally based

reporters cite analysts less and tend to write articles without citing an information intermediary

as a source. This provides evidence of informational advantage for local reporters.

33
 This paper provides evidence regarding the role of the press as a watchdog for accounting

fraud. It is a beginning towards establishing a greater understanding of the press in general.

However, the work should be viewed with several caveats. First, the role as an accounting

watchdog is only one of many roles the press plays in the capital markets. Future work should

address issues such as how the press disseminates already public information, the role of the

press in analyzing a fraud after the initial public announcement as well as the many potential

economic impacts of press coverage. Second, my sample relies on firms being identified as

fraudulent by the SEC and having an AAER written regarding the fraud. These are likely

extreme cases of fraud, and thus readers should be careful in attempting to generalize my results

to the population as a whole. Third, the SEC likely investigates firms that have been identified

in the press as engaging in some form of accounting fraud. In that case, the proportion of firms

“caught” by the press may be overstated. That is, many other frauds may have occurred and

remain undetected by either the press or the SEC. Again, this suggests readers should use

caution when trying to generalize to a fuller sample.

34
 REFERRENCES

Barton, J. “Who Cares About Auditor Reputation” Working Paper. Emory University – June
(2003).

Behr, P., “Looking for Answers in the Enron Story” Neiman Reports Summer (2002): 8-9.

Beneish, M.D., “Incentives and Penalties Related to Earnings Overstatements that Violate
GAAP” The Accounting Review (1999): 425-457.

Beneish, M.D., “The Detection of Earnings Manipulation” Financial Analysts Journal (1999):
24-36.

Bonner, S., Z. Palmrose and S. Young, “Fraud Type and Auditor Litigation: An Analysis of
SEC Accounting and Auditing Enforcement Releases” Accounting Review (1998):
503-532.

Bradshaw, M., B. Bushee and G. Miller, “Accounting Choice, Home Bias, and US Investment in
Non-US Firms”. Working Paper. Harvard University – April (2003).

Bushman, R., J. Piotroski and A. Smith, “What Determines Corporate Transparency?”. Working
Paper. University of North Carolina – Chapel Hill (2003).

Cutler, D., J. Poterba and L. Summers, “What Moves Stock Prices” The Journal of Portfolio
Management – Spring (1989): 4-12.

DeAngelo, H., L. DeAngelo and S. Gilson, “The Collapse of First Executive Corporation Junk
Bonds, Adverse Publicity,” and the “Run on the Bank Phenomenon” Journal of Financial
Economics (1994): 288-336.

DeAngelo, H., L. DeAngelo and S. Gilson, “Perceptions and the Politics of Finance: Junk Bonds
and the Regulatory Seizure of First Capital Life” Journal of Financial Economics (1996):
475-511.

Dechow, P., R. Sloan and A. Sweeney, “Causes and Consequences of Earnings Manipulation:
An Analysis of Firms Subject to Enforcement Actions by the SEC” Contemporary
Accounting Research Vol. 13 No.1 – Spring (1996): 1-36.

DeFond, M. and B. Smith, “Discussion of The Financial and Market Effects of the SEC’s
Accounting and Auditing Enforcement Releases” Journal of Accounting Research
Volume 29 – Supplement. (1991): 143-148.

Desai, H. and P. Jain, “Long-Run Common Stock Returns Following Financial Analysis by
Abraham Briloff”. Working Paper. Southern Methodist University (2000).

35
Djankov, S., C. McLiesh, T. Nenova and A. Shleifer, “Who Owns the Media?”. Working Paper.
Harvard University (2002).

Dyck, A. and L. Zingales, “The Corporate Governance Role of the Media”. Working Paper,
Harvard University (2002).

Feroz, E., K. Park and V. Pastena, “The Financial and Market Effects of the SEC’s Accounting
and Auditing Enforcement Releases” Journal of Accounting Research Volume 29 –
Supplement. (1991): 107-142.

Foster, G., “ Briloff and the Capital Market” Journal of Accounting Research Volume 17 –
Spring (1979): 262-274.

Foster, G., “Rambo IX: Briloff and the Capital Market” Journal of Accounting, Auditing and
Finance Fall – (1987): 409-430.

Frost, C., E. Gordon and A. Hayes, “Stock Exchange Disclosure and Market Liquidity: An
Analysis of 50 International Exchanges”. Working Paper. Dartmouth (2001).

Groseclose, T. and J. Milyo, “A Measure of Media Bias”. Working Paper, University of Chicago
(2003).

Haw, I., H. Bingbing, L. Hwang and W. Wu, “Ultimate Ownership, Income Management, and
Legal and Extra-Legal Institutions”. Working Paper. Texas Christian University (2003).

Herman, E., “The Media and Market in the United States” The Right to Tell – The Role of Mass
Media in Economic Development The World Bank Institute (2002): 61-81.

Islam, R., “Into the Looking Glass: What the Media Tell and Why – an Overview” The Right to
Tell – The Role of Mass Media in Economic Development The World Bank Institute
(2002): 1-23.

Jamieson, K., and K. Campbell, “The Interplay of Influence” Wadsworth, 5th Edition (2001).

Jensen, M., “Toward a Theory of the Press” Economics and Social Institutions, Karl Brunner
Editor. (1979).

Joe, J., “Why Press Coverage of a Client Influences the Audit Opinion” Journal of Accounting
Research Volume 41 – March (2003): 109-133.

Kurtz, H., “The Fortune Tellers: Inside Wall Street’s Game of Money, Media, and Manipulation”
New York: Touchstone, 1st Edition, (2001).

Lewin, G., “The Watchdog Role Business Reporters Need to Play” Neiman Reports Summer
(2002): 19-21.

36
Madrick, J., “A Good Story Isn’t Always the Right One to Tell” Neiman Reports Summer
(2002): 6-7.

McNair, J. “Corporations Work Hard to Prevent Reporting” Neiman Reports Summer (2002):
12-14.

Mullainathan, S. and A. Shleifer, “Media Bias”. Working Paper. Harvard University (2002).

Niederhoffer, V., “The Analysis of World Events and Stock Prices” The Journal of Business
Volume 44 – April (1971): 193-219.

Nunnally, J., Psychometric Theory, New York: McGraw-Hill Companies (1978).

Owen, B., “Media as Industry: Economic Foundations of Mass Communications” The Right to
Tell – The Role of Mass Media in Economic Development The World Bank Institute
(2002): 61-81.

Reuter, J. and E. Zitzewitz, “Do Ads Influence Editors? Advertising and Content Biases in the
Financial Media”. Working Paper. University of Oregon – October (2003).

Serrin, J. and W. Serrin, “Muckraking! The Journalism that Changed America” The New Press
(2002).

Steiger, P. “Not Every Jounalist ‘Missed” the Enron Story” Neiman Reports Summer (2002): 10-
12.

Smith, R. and J. Emshwiller “24 Days” New York: HarperCollins, 1st edition (2003).

Stromberg, D., “Mass Media Competition, Political Competition, and Public Policy”. Working
Paper. Institute for International Economic Studies (2002).

Thompson, R., C. Olsen and J. Dietrich, “Attributes of News About Firms: An Analysis of Firm-
Specific News Reported in the Wall Street Journal Index” Journal of Accounting
Research Volume 25 – Autumn (1987): 245-273.

Wasik, B., “Dismal Beat – the march of personal finance journalism” Harpers – March (2003):
81-87.

Zingales, L., “In Search of New Foundations” The Journal of Finance Volume LV, No. 4 –
August (2000): 1623-1653.

37
 Figure 1

TIME LINE

Infraction Period
INTENSITY SEARCH
Press and Managerial
Company SEC
CAUGHT SEARCH Announces
Between commencement Investigation announces
of activity and
announcement by Auditing and
company

NOTE: Dotted line event “Company Announces Investigation” might not occur.

38
 Table 1
Accounting and Auditing Enforcement Release Firms by End of Year of Fraud

Percentage of Cumulative
Year Number of Firms Sample Percentage
1987 1 .4 .4

1988 1 .4 .8

1989 2 .8 1.6

1990 6 2.3 3.9

1991 7 2.7 6.6

1992 21 8.0 14.6

1993 29 11.0 25.6

1994 28 10.6 36.2

1995 22 8.4 44.6

1996 30 11.4 56.0

1997 32 12.2 68.2

1998 27 10.3 78.5

1999 22 8.4 86.9

2000 24 9.0 95.9

2001 9 3.3 99.2

2002 2 .8 100

Total 263 100 100

A firm is included in the sample if an Accounting and Auditing Enforcement Release was identified on the SEC
website and that release is related to an accounting fraud perpetrated by the firm.

39
 Table 2
Descriptives of Accounting, Audit and Enforcement Releases

Panel A: Length of Violations in Years

Number of
Years of Percentage of Cumulative
Violation Number of Firms Sample Percentage

1 79 30.0 30.0

2 93 35.4 65.4

3 44 16.7 82.1

4 24 9.1 91.2

5 11 4.2 95.5

6 7 2.7 98.2

7 5 1.8 100

Total 263 100 100

Panel B: Number of Violations

Number of
Violations
Reported Percentage of Cumulative
Per Firm Number of Firms Sample Percentage

1 41 15.6 15.6

2 88 33.5 49.1

3 66 25.1 74.1

4 34 12.9 87.0

5 22 8.4 95.4

6 10 3.8 99.2

7 2 .8 100

Total 263 100 100

40
Panel C: Type of Violations

Type of Violation Number of Firms Percentage of Sample

Revenue Manipulation 129 49.1

Asset Overstatement 93 35.4

Materially Misleading
Public Statements 69 26.3

Reserve Manipulation 47 17.9

Misappropriation of
Funds 63 24.0

Accounting and Auditing Enforcement Releases (AAER) report the length of the violation is months. However, for
the purposes of Panel A in this table the violations are reported in years based on rounding months up to the next
highest year. The number of violations is calculated based on the AAER. In some cases the SEC issues more than
one AAER for a single instances of fraud (generally due to multiple users). In that case, all unique violations are
included. The types of violation are collected from the AAER. As is obvious from Panel B, most AAER include
more than one violation. Thus, a firm may be included in more than one of the categories shown in Panel C.

41
 Table 3
Descriptives of Press Coverage and Managerial Press Releases

Number Number of Firms Number of Firms

Provided per with Articles by Percentage with Managerial Percentage
Month Press of Sample Press Releases of Sample

(0,1) 42 16.0 111 42.2

[1,2) 29 11.0 71 27.0

[2,5) 54 20.5 58 22.1

[5,10) 39 14.8 14 5.3

[10,20) 40 15.2 7 2.7

[20,50) 27 10.3 1 .4

[50,100) 14 5.3 1 .4

≥ 100 18 6.8 0 0

Total 263 100 263 100

The number of each item presented per month is calculated by searching Factiva over the entire period of the
accounting fraud and deflating the total number of items found by the length of the fraud (in months). Thus, it
represents the average quantity that occurs in a month. Articles by the press were identified by performing a full-
text search of all Factiva sources except for PR Newswire, Business Wire and several summary features that only
record a name and the fact that some event occurred. Managerial press releases were identified by searching PR
Newswire and Business Wire. Some form of the company name was required to appear in the heading or first
paragraph of the press release.

42
 Table 4
Firms “Caught” by Press Article presented by Last Year of Accounting Fraud

Number of
Firms with Percentage of
“Caught” Current Year with
Year Number of Firms Article “Caught” Article
1987 1 1 100

1988 1 0 0

1989 2 0 0

1990 6 1 16.7

1991 7 1 14.3

1992 21 4 19.1

1993 29 9 31.0

1994 28 12 42.9

1995 22 10 45.5

1996 30 9 30.0

1997 32 8 25.0

1998 27 6 22.2

1999 22 5 22.8

2000 24 6 25.0

2001 9 3 33.3

2002 2 1 50.0

Total 263 76 na

Percent of
Total 100% 28.9%

A firm is included in the sample if an Accounting and Auditing Enforcement Release was identified on the SEC
website and that release is related to an accounting fraud perpetrated by the firm. A firm is considered “caught” by
the press if an article appears questioning the firm’s accounting prior to public disclosure by the firm (or SEC) that
an accounting problem exists.

43
 Table 5
Cross-Sectional Examination of Characteristics of Firms and Fraud that Lead to an Article by the
Press

Panel A: Means and Medians

P-value Of
Variable Caught Not Caught Difference
PRESSINT Mean 51.23 24.21 .032
Median 8.36 4.96 .048
ANALYST Mean .34 .22 .019
Median 1 0 .019
MV Mean 4.16 2.06 .122
Median 2.96 1.09 .034
BIGADS Mean .1447 .1765 .267
Median 0 0 .266
NUMVIOL Mean 3.01 2.7 .063
Median 3 3 .175
LENGTH Mean 26.01 23.87 .205
Median 24 21 .107
MISLEAD Mean .37 .22 .009
Median 0 0 .006
STEAL Mean .33 .20 .022
Median 0 0 .015
REV Mean .51 .48 .325
Median 1 0 .320
ASSETOVER Mean .26 .39 .051
Median 0 0 .025

44
 Table 5 (Continued)
Cross-Sectional Examination of Characteristics of Firms and Fraud that Lead to an Article by the
Press

Panel B: Logit Analyses (P-values in Parentheses below Coefficients)

Caught = α t + β 1t InformationVariable + β 2 NUMVIOL + β 3 LENGTH + β 4 MISLEAD + β 5 STEAL +

β 6 REV + β 7 ASSETOVER + ε

INFORMATON VARIABLE
PRESSINT ANALYST MV
INTERCEPT -1.4406 -1.4882 -1.0537
(.0030) (.0002) (.0262)

Info. Variable .0029 .6177 .00000019

(.0373) (.02585) (.1226)

BIGAD .0760 .0177 .2496

(.4251) (.4823) (.2986)

NUMVIOL .0223 .0331 -.0173

(.4375) (.4072) (.4611)

LENGTH .0046 .0061 -.0055

(.2835) (.2229) (.3069)

MISLEAD .6239 .6478 .8397

(.0366) (.0320) (.0292)

STEAL .5986 .5833 .7726

(.0517) (.0566) (.0451)

REV .1267 -.0217 .0754

(.7030) (.9459) (.8446)

ASSETOVER -.4639 -.4978 .2102

(.2030) (.1675) (.6373)

Likelihood 318.25 318.25 224.31

Ratio (chi) (17.12) (17.52) (12.02)

(two-tailed p-values for INTERCEPT, REV and ASSETOVER, one-tailed for other variables).

This table presents analyses of differences between firms for which fraud was identified (CAUGHT=1) vs. those
which were not (CAUGHT=0). Sample size for all analyses other than those including MV is 263. Seventy-six
firms are coded as CAUGHT=1. Sample size for analyses with MV is 181. Fifty-five of those firms are coded as
CAUGHT=1. PRESSINT is measured as the total number of articles by the press over the fraud violation period
divided by the months of the violation. Articles were identified using Factiva. ANALYST is an indicator variable
coded as 1 if the firm is in the top quartile of analyst following for the sample, 0 otherwise. Analyst following is
measured at the last month of the violation period and is from IBES. MV is the market value of the firm in billions.
Market value is measured on the last day of the violation period and is obtained from CRSP. BIGAD is an indicator
variable coded as 1 if the firm is in an industry that was in Advertising Age’s top 15 advertisers for 1985, 1990,
45
1995, and 2000. Otherwise, the coding is 0. NUMVIOL is the number of violations cited in the AAER. LENGTH
is the length of the violation as documented in the AAER. MISLEAD is an indicator variable coded as one if the
violations on the AAER included a materially misleading public statement or report, 0 otherwise. STEAL is an
indicator variables coded as 1 if the AAER indicates management misappropriated funds as a portion of the fraud, 0
otherwise. REV is an indicator variable coded as 1 if the AAER violations included revenue manipulation, 0
otherwise. ASSETOVER is an indicator variable coded as 1 if the AAER violations included revenue manipulation,
0 otherwise.

46
 Table 6
Types of Information Sources Used by the Press

Information Percentage of Cumulative

Source Number of Articles Sample Percentage

None Given 14 18.4 18.4

Analyst 23 30.3 48.7

Legal Cases 15 19.7 68.4

Auditor
Resignation 12 15.8 84.2

SEC filing 4 5.3 89.5

Customer 3 4.0 93.5

Industry
Insider 2 2.6 96.1

Anonymous
Tipster 2 2.6 98.7

Web 1 1.3 1.3

Total 76 100 100

Source is based on the source that is attributed as providing the primary information or to have provided the
information that initiated the article.

47
 Table 7
The Relation among Types of Information Sources and Characteristics of the Publications
Providing Articles

Type of Percentage of Cumulative

Publication Number of Articles Sample Percentage

National
Business 29 38.2 38.2

Local
Market 23 30.3 68.5

Electronic
Business 15 19.7 88.2

Trade
Publications 7 9.2 97.4

National
Non-
Business 2 2.6 100

Total 76 100 100

National Business publications are publications that cover the nation (or global) areas. This category includes the
Wall Street Journal, Fortune, etc. Local Market publications are generally regional papers, such as The San
Francisco Chronicle, Chicago-Sun Times etc. Electronic Business publications consist almost entirely of the Dow
Jones News Service with one article from Bloomberg. Trade publications are based on covering a specific industry
in depth. Examples include Boating Industry and Business Insurance. The National non-business publications are
USA Today and The New York Times.

48
 Table 8
Relation between Sources of Information and Characteristics of the Publications

Table A: Relation between Sources of Information and Type of Publication

Information
Source National Business Local Market Electronic Business Trade Publications

None Given 8 6 0 0

Analyst 12 4 6 1

Legal Cases 2 8 4 1

Auditor
Resignation 3 5 2 2

Other 4 2 3 3

Table B: Relation between Sources of Information and Region of Publication or By-line

Percentage of
Information Regional Percentage of
Source Regional Sample Non-Regional Non-Regional Sample

None Given 8 33.3 6 11.5

Analyst 4 16.7 19 36.5

Legal Cases 4 16.7 11 21.2

Auditor
Resignation 5 20.8 7 13.5

Other 3 12.5 9 17.3

Total 24 100 52 100

Source is based on the source that is attributed as providing the primary information or to have initiated for the
article. National Business publications are publications that cover the nation (or global) areas. This category
includes the Wall Street Journal, Fortune, etc. Local Market publications are generally regional papers, such as The
San Francisco Chronicle, Chicago-Sun Times etc. However, for the purposes of this table they also include the two
national non-business publications (The New York Times and USA Today). Electronic Business publications consist
almost entirely of the Dow Jones News Service with one article from Bloomberg. Trade publications are based on
covering a specific industry in depth. Examples include Boating Industry and Business Insurance. An article is
considered regional if it is from a local publication in the same region or a national publication but provides a by-
line indicating the story was written locally.

49