Personal Information: Data Breaches Are Frequent, but Evidence of
Resulting Identity Theft Is Limited; However, the Full Extent Is 
Unknown (04-JUN-07, GAO-07-737).				 
                                                                 
In recent years, many entities in the private, public, and	 
government sectors have reported the loss or theft of sensitive  
personal information. These breaches have raised concerns in part
because they can result in identity theft--either account fraud  
(such as misuse of credit card numbers) or unauthorized creation 
of new accounts (such as opening a credit card in someone else's 
name). Many states have enacted laws requiring entities that	 
experience breaches to notify affected individuals, and Congress 
is considering legislation that would establish a national breach
notification requirement. GAO was asked to examine (1) the	 
incidence and circumstances of breaches of sensitive personal	 
information; (2) the extent to which such breaches have resulted 
in identity theft; and (3) the potential benefits, costs, and	 
challenges associated with breach notification requirements. To  
address these objectives, GAO reviewed available reports on data 
breaches, analyzed 24 large data breaches, and gathered 	 
information from federal and state government agencies, 	 
researchers, consumer advocates, and others.			 
-------------------------Indexing Terms------------------------- 
REPORTNUM:   GAO-07-737 					        
    ACCNO:   A71532						        
  TITLE:     Personal Information: Data Breaches Are Frequent, but    
Evidence of Resulting Identity Theft Is Limited; However, the	 
Full Extent Is Unknown						 
     DATE:   06/04/2007 
  SUBJECT:   Computer security					 
	     Cost analysis					 
	     Data collection					 
	     Databases						 
	     Fraud						 
	     Government information				 
	     Identity theft					 
	     Information management				 
	     Information security				 
	     Internal controls					 
	     Reporting requirements				 
	     Strategic planning 				 

******************************************************************
** This file contains an ASCII representation of the text of a  **
** GAO Product.                                                 **
**                                                              **
** No attempt has been made to display graphic images, although **
** figure captions are reproduced.  Tables are included, but    **
** may not resemble those in the printed version.               **
**                                                              **
** Please see the PDF (Portable Document Format) file, when     **
** available, for a complete electronic file of the printed     **
** document's contents.                                         **
**                                                              **
******************************************************************
GAO-07-737

   

     * [1]Results in Brief
     * [2]Background
     * [3]Available Evidence Indicates That Data Breaches Occur Freque

          * [4]Several Sources Indicate That Breaches of Sensitive Personal

               * [5]Media Reports
               * [6]Federal Law Enforcement Agencies
               * [7]House Government Reform Committee and DHS
               * [8]Federal Banking Regulators
               * [9]State Agencies
               * [10]Other Sources

          * [11]Source, Cause, Size, and Content of Breaches Have Varied Wid

               * [12]Type of Entity
               * [13]Cause of Breach
               * [14]Number of Records Breached
               * [15]Types of Data Breached

     * [16]Consequences of Data Breaches Are Not Fully Known, but Clear

          * [17]Federal Law Enforcement Agencies and Industry Associations I
          * [18]Of 24 Large Publicly Reported Breaches, 4 Apparently Resulte
          * [19]Challenges Exist in Determining the Link between Data Breach
          * [20]Type of Data Compromised and Other Factors Influence Potenti

     * [21]Breach Notification Requirements Can Serve to Encourage Bett

          * [22]Notification Requirements May Create Incentives for Improved
          * [23]Breach Notification Requirements Present a Variety of Potent
          * [24]Challenges Exist in Complying with and Developing Breach Not

               * [25]Complying with Notification Requirements
               * [26]Setting an Appropriate Notification Standard

     * [27]Agency Comments
     * [28]GAO Contact
     * [29]Staff Acknowledgments
     * [30]GAO's Mission
     * [31]Obtaining Copies of GAO Reports and Testimony

          * [32]Order by Mail or Phone

     * [33]To Report Fraud, Waste, and Abuse in Federal Programs
     * [34]Congressional Relations
     * [35]Public Affairs

Report to Congressional Requesters

United States Government Accountability Office

GAO

June 2007

PERSONAL INFORMATION

Data Breaches Are Frequent, but Evidence of Resulting Identity Theft Is
Limited; However, the Full Extent Is Unknown

GAO-07-737

Contents

Letter 1

Results in Brief 5
Background 7
Available Evidence Indicates That Data Breaches Occur Frequently and Under
Varying Circumstances 10
Consequences of Data Breaches Are Not Fully Known, but Clear Evidence of
Identity Theft Has Been Found in Relatively Few Breaches 21
Breach Notification Requirements Can Serve to Encourage Better Data
Security Practices and Alert Consumers, but They Also Present Costs and
Challenges 31
Agency Comments 40
Appendix I Scope and Methodology 42
Appendix II GAO Contact and Staff Acknowledgments 45
GAO Contact 45
Staff Acknowledgments 45

Table

Table 1: Twenty-Four Large Publicly Reported Data Breaches and Evidence of
Resulting Identity Theft, January 2000 - June 2005 26

Figure

Figure 1: Application of Notification Standards under Different Breach
Scenarios 37

This is a work of the U.S. government and is not subject to copyright
protection in the United States. It may be reproduced and distributed in
its entirety without further permission from GAO. However, because this
work may contain copyrighted images or other material, permission from the
copyright holder may be necessary if you wish to reproduce this material
separately.

Abbreviations

DHS Department of Homeland Security
FBI Federal Bureau of Investigation
FDIC Federal Deposit Insurance Corporation
FTC Federal Trade Commission
SSN Social Security number
USPIS United States Postal Inspection Service
VA Department of Veterans Affairs

United States Government Accountability Office
Washington, DC 20548

June 4, 2007

The Honorable Spencer Bachus
Ranking Member
Committee on Financial Services
House of Representatives

The Honorable Michael N. Castle
House of Representatives

The Honorable Darlene Hooley
House of Representatives

The Honorable Steven C. LaTourette
House of Representatives
The Honorable Dennis Moore
House of Representatives

As a result of advances in computer
technology and electronic storage, many different sectors and entities now
maintain electronic records containing vast amounts of personal
information on virtually all American consumers. In recent years, a number
of entities--including financial service firms, retailers, universities,
and government agencies--have collectively reported the loss or theft of
large amounts of sensitive personal information. Some of these data
breaches--such as those involving TJX Companies and the Department of
Veterans Affairs (VA)--have received considerable publicity and have
highlighted concerns about the protections afforded sensitive personal
information.1 Policymakers, consumer advocates, and others have raised
concerns that data breaches can contribute to identity theft, in which an
individual's sensitive personal The Honorable Spencer Bachus Ranking
Member Committee on Financial Services House of Representatives The
Honorable Michael N. Castle House of Representatives The Honorable Darlene
Hooley House of Representatives The Honorable Steven C. LaTourette House
of Representatives The Honorable Dennis Moore House of Representatives As
a result of advances in computer technology and electronic storage, many
different sectors and entities now maintain electronic records containing
vast amounts of personal information on virtually all American consumers.
In recent years, a number of entities--including financial service firms,
retailers, universities, and government agencies--have collectively
reported the loss or theft of large amounts of sensitive personal
information. Some of these data breaches--such as those involving TJX
Companies and the Department of Veterans Affairs (VA)--have received
considerable publicity and have highlighted concerns about the protections
afforded sensitive personal information.1 Policymakers, consumer
advocates, and others have raised concerns that data breaches can
contribute to identity theft, in which an individual's sensitive personal
information is used fraudulently. The Federal Trade Commission (FTC),
which is responsible for taking complaints from victims and sharing them
with law enforcement agencies, has noted that identity theft is a serious
problem--millions of Americans are affected each year, and victims may
face substantial costs and time to repair the damage to their good name
and credit record.

1In January 2007, The TJX Companies, Inc., publicly disclosed a data
breach that compromised sensitive personal information, including credit
and debit card data, associated with more than 45 million customer
accounts. In May 2006, VA reported that computer equipment containing
sensitive personal information on approximately 26.5 million veterans and
active duty members of the military was stolen from the home of a VA
employee. The equipment was eventually recovered, and forensic analysts
concluded that it was unlikely that the personal information contained
therein was compromised. See GAO, Privacy: Lessons Learned About Data
Breach Notification, [36]GAO-07-657 (Washington, D.C.: Apr. 30, 2007).

Although there is no commonly agreed-upon definition, the term "data
breach" generally refers to an organization's unauthorized or
unintentional exposure, disclosure, or loss of sensitive personal
information, which can include personally identifiable information such as
Social Security numbers (SSN) or financial information such as credit card
numbers.2 Data breaches can take many forms and do not necessarily lead to
identity theft. The term "identity theft" is broad and encompasses many
types of criminal activities, including fraud on existing accounts--such
as unauthorized use of a stolen credit card number--or fraudulent creation
of new accounts--such as using stolen data to open a credit card account
in someone else's name. Depending on the type of information compromised
and how it is misused, identity theft victims can face a range of
potential harm, from the inconvenience of having a credit card reissued to
substantial financial losses and damaged credit ratings.

Beginning with California in 2002, at least 36 states have enacted breach
notification laws--that is, laws that require certain entities that
experience a data breach to notify individuals whose personal information
was lost or stolen. There is no federal statute that requires most
companies or other entities to notify affected individuals of data
breaches, although federal banking regulatory agencies have issued
guidance on breach notification to the banks, thrifts, and credit unions
they supervise.3 In addition, the Office of Management and Budget has
issued guidance--developed by the President's Identity Theft Task
Force--on responding to data breaches at federal agencies.4 Because a
number of bills have been introduced in Congress that would establish a
national breach notification requirement, you asked us to review the costs
and benefits of such a requirement and the link between data breaches and
identity theft.5 As agreed with your offices, this report examines (1)
what is known about the incidence and circumstances of breaches of
sensitive personal information; (2) what information exists on the extent
to which breaches of sensitive personal information have resulted in
identity theft; and (3) the potential benefits, costs, and challenges
associated with breach notification requirements.

2In this report we use "personally identifiable information" to refer to
any information that can be used to distinguish or trace an individual's
identity--such as name, Social Security number, driver's license number,
and mother's maiden name--because such information generally may be used
to establish new accounts, but not to refer to other "means of
identification," as defined in 18 U.S.C. S 1028(7), including account
information such as credit or debit card numbers.

This report focuses on breaches of sensitive personal data that can be
used to commit identity theft, and not on breaches of other sensitive
data, such as medical records or proprietary business information. To
address the first two objectives, we obtained and analyzed information on
data breaches that have been reported in the media and aggregated by three
private research and advocacy organizations, as well as information on
breaches collected by state agencies in New York and North Carolina,
federal banking regulators, and federal law enforcement agencies.6 We also
collected information on breaches experienced by federal agencies compiled
by the House Government Reform Committee in 2006 and by the Department of
Homeland Security (DHS).7 In addition, we conducted a literature search of
relevant articles, reports, and studies. We also conducted interviews
with, and obtained documents from, representatives of federal agencies,
including the FTC, the Department of Justice, DHS, and the federal banking
regulatory agencies; selected state government agencies and the National
Association of Attorneys General; private and nonprofit research
organizations; and consumer protection and privacy advocacy groups.
Further, we obtained information from industry and trade associations
representing key sectors--including financial services, retail sales,
higher education, health care, and information services--that have
experienced data breaches. In addition, for the second objective, we
examined the 24 largest (in terms of number of records breached) data
breaches reported by the news media from January 2000 through June 2005
and tracked by private groups. For each of these breaches, we reviewed
media reports and other publicly available information, and conducted
interviews, where possible, with representatives of the entities that
experienced the breaches, in an attempt to identify any known instances of
identity theft that resulted from the breaches. We also examined five
breaches that involved federal agencies, which were selected because they
represented a variety of different circumstances. For the third objective,
we reviewed the federal banking regulatory agencies' proposed and final
guidance related to breach notification, and interviewed representatives
of each agency regarding their consideration of potential costs, benefits,
and challenges during development of the guidance. Further, we reviewed
the strategic plan and other documents issued by the President's Identity
Theft Task Force. In addition, we conducted a review of the effects of
California's breach notification law, which included interviewing and
gathering information from California state officials and selected
California companies, educational institutions, and other entities subject
to the law's notification requirements.

3See Interagency Guidance on Response Programs for Unauthorized Access to
Customer Information and Customer Notice, 70 Fed. Reg. 15736 (Mar. 29,
2005). The five federal banking regulatory agencies are the Office of the
Comptroller of the Currency, the Office of Thrift Supervision, the Board
of Governors of the Federal Reserve System, the Federal Deposit Insurance
Corporation, and the National Credit Union Administration. The National
Credit Union Administration issued its guidance (which was substantially
identical) separately from the other four regulators (see Security Program
and Appendix B--Guidance on Response Programs for Unauthorized Access to
Member Information and Member Notice, 70 Fed. Reg. 22764 (May 2, 2005)).

4The President's Identity Theft Task Force--chaired by the Attorney
General and cochaired by the Chairman of the Federal Trade Commission and
comprising 17 federal agencies and departments--was charged with
developing a comprehensive national strategy to combat identity theft.
Exec. Order No. 13,402, Strengthening Federal Efforts to Protect Against
Identity Theft, 71 Fed. Reg. 27945 (May 10, 2006). The task force's
guidance was distributed in a memorandum from the Office of Management and
Budget to the heads of federal agencies and departments. See Office of
Management and Budget Memorandum for the Heads of Departments and
Agencies, Recommendations for Identity Theft Related Data Breach
Notification, Sept. 20, 2006. In May 2007, the Office of Management and
Budget issued a memorandum that updated the September 2006 guidance and,
among other things, required agencies to develop and implement breach
notification policies within 120 days. See Office of Management and Budget
Memorandum for the Heads of Executive Departments and Agencies,
Safeguarding Against and Responding to the Breach of Personally
Identifiable Information, M-07-16 (May 22, 2007).

5See, for example, Data Security Act of 2007, H.R. 1685, 110th Cong.
(2007); Notification of Risk to Personal Data Act of 2007, S. 239, 110th
Cong. (2007); Personal Data Privacy and Security Act of 2007, S. 495,
110th Cong. (2007); Data Accountability and Trust Act, H.R. 958, 110th
Cong. (2007); and Identity Theft Prevention Act, S. 1178, 110th Cong.
(2007).

6The three private organizations are Attrition, Identity Theft Resource
Center, and Privacy Rights Clearinghouse. We reviewed data on breaches in
New York and North Carolina because they represent two large states that
maintain centralized information on data breaches.

7The House Government Reform Committee was renamed the House Oversight and
Government Reform Committee in the 110th Congress.

We conducted our review from August 2006 through April 2007 in accordance
with generally accepted government auditing standards. A more extensive
discussion of our scope and methodology appears in appendix I.

Results in Brief

While comprehensive data do not exist, available evidence suggests that
breaches of sensitive personal information have occurred frequently and
under widely varying circumstances. For example, more than 570 data
breaches have been reported in the news media from January 2005 through
December 2006, according to our analysis of lists maintained by three
private organizations that track such breaches. Further, a House
Government Reform Committee survey of federal agencies identified more
than 788 data breaches at 17 agencies from January 2003 through July 2006.
Of the roughly 17,000 federally supervised banks, thrifts, and credit
unions, several hundred have reported data breaches to their federal
regulators over the past 2 years. In addition, officials in New York
State--which requires public and private entities to report data breaches
to a centralized source--reported receiving notice of 225 breaches from
December 7, 2005, through October 5, 2006. Data breaches have occurred
across a wide range of entities, including federal, state, and local
government agencies; retailers; financial institutions; colleges and
universities; and medical facilities. Some studies indicate that most
publicly reported breaches resulted from intentional actions, such as a
stolen laptop computer, rather than accidental occurrences, such as a lost
laptop computer, but this may be because breaches related to criminal
activity are perhaps more likely to be reported. Media-reported breaches
have varied significantly in size, ranging from 10 records to tens of
millions of records. Most of these breaches have compromised data that
included personally identifiable information, while others have involved
only account information such as credit card numbers.

The extent to which data breaches result in identity theft is not well
known, in large part because it can be difficult to determine the source
of the data used to commit identity theft. Although we identified several
cases where breaches reportedly have resulted in identity theft--that is,
account fraud or unauthorized creation of new accounts--available data and
interviews with researchers, law enforcement officials, and industry
representatives indicated that most breaches have not resulted in detected
incidents of identity theft. For example, our review of the 24 largest
breaches that appeared in the news media from January 2000 through June
2005 found that 3 breaches appeared to have resulted in fraud on existing
accounts, and 1 breach appeared to have resulted in the unauthorized
creation of new accounts. For 18 of the breaches, no clear evidence had
been uncovered linking them to identity theft; and for the remaining 2, we
did not have sufficient information to make a determination. Determining
the link between data breaches and identity theft is challenging,
primarily because identity theft victims often do not know how their
personal information was obtained, and it may be up to a year or more
before stolen data are used to commit a crime. Some studies by private
researchers have found little linkage between data breaches and identity
theft, although our review found these studies had methodological
limitations. Finally, the circumstances of a breach can greatly affect the
potential harm that can result. For example, unauthorized creation of new
accounts generally can occur only when a breach includes personally
identifiable information. Further, breaches that are the result of
intentional acts generally are considered to pose more risk than
accidental breaches, according to federal officials.

Requiring consumer notification of data breaches may encourage better data
security practices and help deter or mitigate harm from identity theft,
but it also involves monetary costs and challenges such as determining an
appropriate notification standard. Representatives of federal banking
regulators, other government agencies, industry associations, and other
affected parties told us that breach notification requirements have
encouraged companies and other entities to improve their data security
practices to minimize legal liability or avoid public relations risks that
may result from a publicized breach of customer data. Further, notifying
affected consumers of a breach gives them the opportunity to mitigate
potential risk--for example, by reviewing their credit card statements and
credit reports, or placing a fraud alert on their credit files. Some
privacy advocates and others have noted that even when the risk of actual
financial harm is low, breach notification is still important because
individuals have a basic right to know how their personal information is
being handled and when it has been compromised. At the same time, affected
entities incur monetary costs to comply with notification requirements.
For example, 31 companies that responded to a 2006 survey said they
incurred an average of $1.4 million per breach, for costs such as mailing
notification letters, call center expenses, courtesy discounts or
services, and legal fees. In addition, organizations subject to
notification requirements told us they face several challenges, including
the lack of clarity in some state statutes about when a notification is
required, difficulty identifying and locating affected individuals, and
difficulty complying with varying state requirements. Notification
standards--that is, the circumstances surrounding a data breach that
"trigger" the required notification--vary among the states. Some parties,
such as the National Association of Attorneys General, have advocated that
a breach notification requirement should apply broadly in order to give
consumers a greater level of protection and because the risk of harm is
not always known. The guidance provided by federal banking regulators lays
out a more risk-based approach, aimed at ensuring that affected
individuals receive notices only when they are at risk of identity theft
or other related harm. Such an approach was also adopted by the
President's Identity Theft Task Force, which recommended a risk-based
standard for breach notification applicable to both government agencies
and private entities. As we have noted in the past, care is needed in
defining appropriate criteria for incidents that merit notification.
Should Congress choose to enact a federal breach notification requirement,
use of such a risk-based standard could avoid undue burden on
organizations and unnecessary and counterproductive notifications of
breaches that present little risk.

This report contains no recommendations. We provided a draft of this
report to FTC and provided selected portions of the draft to federal
banking regulatory agencies and relevant federal law enforcement agencies.
These agencies provided technical comments, which we have incorporated in
this report as appropriate.

Background

Breaches of sensitive personal data in recent years at companies,
universities, government agencies, and other organizations have heightened
public awareness about data security and the risks of identity theft, and
have led to the introduction of breach notification requirements in many
state legislatures. As of April 2007, at least 36 states had enacted some
form of law requiring that affected individuals be notified in the event
of a data breach; California's law, enacted in 2002, was the first such
state requirement.8 States' notification requirements vary, particularly
with regard to the applicable notification standard--the event or
circumstance that triggers a required notification. Requirements also vary
in terms of the data to which they apply--for example, some apply to paper
documents as well as electronic records.

8Cal. Civ. Code S 1798.82.

There is currently no federal statute that requires most companies and
other entities that experience a data breach to notify individuals whose
personal information was lost or stolen. However, the Gramm-Leach-Bliley
Act established requirements for federally supervised financial
institutions to safeguard customer information.9 To clarify these
requirements, the federal banking regulators issued interagency guidance
in 2005 to the banks, thrifts, and credit unions they supervise related to
their handling of data breaches. Under this guidance, these institutions
are expected to develop and implement a response program to address
unauthorized access to customer information maintained by the institution
or its service providers; and if they experience a breach, they are to
notify their primary federal regulator as soon as possible and--depending
on the circumstances of the incident--notify their affected customers. In
addition, in September 2006 the President's Identity Theft Task Force
developed guidance for federal agencies on responding to breaches
involving agency data, including the factors to consider in determining
whether to notify affected individuals. The task force released a
strategic plan for combating identity theft in April 2007, which contained
among its recommendations a proposal for establishing a national breach
notification requirement.10 Further, in December 2006, the Department of
Veterans Affairs Information Security Enhancement Act of 2006 became law,
which, among other things, requires VA to prescribe regulations providing
for the notification of data breaches occurring at the department.11 A
number of bills have been introduced in Congress that would more broadly
require companies and other entities to notify individuals when such
breaches occur, and Congress has held several hearings related to data
breaches.12

9Pub. L. No. 106-102, tit. V, subtit. A, 113 Stat. 1338 (Nov. 12, 1999)
(codified at 15 U.S.C. S 6801-6809).

10President's Identity Theft Task Force, Combating Identity Theft: A
Strategic Plan (Washington, D.C.: Apr. 11, 2007).

11Pub. L. No. 109-461, tit. IX, 120 Stat. 3450 (Dec. 22, 2006), codified
at 38 U.S.C. S 5721-5728, 7901-7907.

Identity theft occurs when individuals' identifying information is used
without authorization in an attempt to commit fraud or other crimes.13
There are two primary forms of identity theft. First, identity thieves can
use financial account identifiers, such as credit card or bank account
numbers, to take over an individual's existing accounts to make
unauthorized charges or withdraw money. Second, thieves can use
identifying data, which can include such things as SSNs and driver's
license numbers, to open new financial accounts and incur charges and
credit in an individual's name, without that person's knowledge. This
second form of identity theft is potentially the most damaging because,
among other things, it can take some time before a victim becomes aware of
the problem, and it can cause substantial harm to the victim's credit
rating. While some identity theft victims can resolve their problems
quickly, others face substantial costs and inconvenience repairing damage
to their credit records. According to FTC, millions of Americans have
their identities stolen each year. Roughly 85 percent of these cases
involve the misuse of existing accounts and 35 percent involve new account
creation or other fraud. (Twenty percent of the total involve both.)
Identity thieves obtain sensitive personal information using a variety of
methods. One potential source is a breach at an organization that
maintains large amounts of sensitive personal information. However,
identity theft can also occur as a result of the loss or theft of data
maintained by an individual, such as a lost or stolen wallet or a thief
digging through household trash.

12For example, Identity Theft: Recent Developments Involving the Security
of Sensitive Consumer Information: Hearing Before the Senate Comm. on
Banking, Housing, and Urban Affairs, 109th Cong., 1st Sess. (2005);
Securing Electronic Personal Data: Striking a Balance Between Privacy and
Commercial and Governmental Use: Hearing Before the Senate Comm. on the
Judiciary, 109th Cong., 1st Sess. (2005); Assessing Data Security:
Preventing Breaches and Protecting Sensitive Information: Hearing Before
the House Comm. on Financial Services, 109th Cong., 1st Sess. (2005);
Securing Consumers' Data: Options Following Security Breaches: Hearing
Before the Subcomm. on Commerce, Trade, and Consumer Protection of the
House Comm. on Energy and Commerce, 109th Cong., 1st Sess. (2005).

13For additional information on identity theft, see GAO, Identity Theft:
Some Outreach Efforts to Promote Awareness of New Consumer Rights Are
Under Way, [37]GAO-05-710 (Washington, D.C.: Jun. 30, 2005) and Identity
Theft: Prevalence and Cost Appear to be Growing,  [38]GAO-02-363
(Washington, D.C.: Mar. 1, 2002).

The Identity Theft and Assumption Deterrence Act of 1998 made identity
theft a federal crime and charged FTC with taking complaints from identity
theft victims; sharing these complaints with federal, state, and local law
enforcement agencies; and providing the victims with informational
materials to assist them.14 Because identity theft is typically not a
stand-alone crime but rather a component of one or more crimes such as
bank fraud, credit card fraud, and mail fraud, a number of federal law
enforcement agencies can have a role in investigating identity theft
crimes, including the Federal Bureau of Investigation (FBI), U.S. Postal
Inspection Service (USPIS), U.S. Secret Service (Secret Service), the
Office of the Inspector General of the Social Security Administration, and
Immigration and Customs Enforcement.

Available Evidence Indicates That Data Breaches Occur Frequently and Under
Varying Circumstances

Available evidence from media reports, federal and state agencies, and
private institutions, collectively, suggests that data breaches occur with
some frequency. For example, our analysis of the lists of data breaches
compiled by three private research and advocacy organizations shows more
than 570 breaches reported by the news media from January 2005 through
December 2006. Data breaches have occurred across a range of entities,
including federal, state, and local government agencies; retailers;
financial institutions; colleges and universities; and medical facilities.
Breaches have varied in size and have resulted from both criminal actions
and accidental incidents. Most of the breaches reported in the news media
have involved data that included personal identifiers such as SSNs, while
others have involved only account information such as credit card numbers.

Several Sources Indicate That Breaches of Sensitive Personal Information Are
Frequent

No federal agency or other organization tracks all data breaches, and
definitions of what constitutes a data breach may vary. Although there are
no comprehensive data on the extent of data breaches nationwide,
government officials, trade association representatives, researchers, and
consumer and privacy advocates we interviewed agreed that breaches of
sensitive personal information occur frequently. For example,
representatives of a variety of organizations--including the Department of
Justice, California's Office of Privacy Protection, the Consumer Data
Industry Association (a trade group representing many information
resellers), and the Ponemon Institute (a private research
organization)--characterized data breaches in the United States as being
"prevalent" or "common." Although we did not identify comprehensive data
on the extent of data breaches, available information from several sources
does corroborate the anecdotal evidence that such breaches occur
frequently.15

14Pub. L. No. 105-318, 112 Stat. 3007 (Oct. 30, 1998). In addition to FTC,
other federal agencies maintain data on identity theft. For example, the
Internet Crime Complaint Center, a joint venture of the FBI and the
National White Collar Crime Center, receives Internet-related identity
theft complaints, which it shares with law enforcement agencies throughout
the country.

  Media Reports

Over the past few years, several hundred data breaches have been reported
each year by newspapers and other news media. Three private organizations
that focus on information privacy and security issues--Privacy Rights
Clearinghouse, Identity Theft Resource Center, and Attrition--track data
breaches reported in newspapers, magazines, and other publicly available
sources of news and information.16 Our analysis of the three lists of data
breaches maintained by these organizations indicated that at least 572
breaches were reported in the news media from January 2005 through
December 2006.17 These breaches were reported to have affected more than
80 million records.18 However, for several reasons, these lists likely
understate the true extent of data breaches in the United States. First,
organizations might not voluntarily disclose data breaches that they
experience. Second, some breaches that organizations do disclose may not
appear in the news media, particularly if the breach was limited in scope.
Finally, the three organizations compiling these lists may not have
identified all of the breaches reported in the news media--for example,
many breaches did not appear on all three lists, suggesting that none
represents an exhaustive list of all breaches that have appeared in the
news.

15Because the breaches cited in this section of the report derive from
different sources, there may be some overlap among the numbers cited by
these sources.

16Privacy Rights Clearinghouse is a nonprofit consumer education and
advocacy project whose purpose is to advocate for consumers' privacy
rights in public policy proceedings. Identity Theft Resource Center is a
nonprofit organization that provides consumer and victim support and
advises governmental agencies, legislators, and companies on the issue of
identity theft. Attrition is an information security-related Web site
maintained by volunteers.

17Representatives of these three organizations indicated that their
definition of a data breach was consistent with the definition used in
this report. However, we did not independently confirm whether the
individual breaches reported by the media and tracked by these groups met
the criteria for this definition, and it is possible that some of them do
not. We reviewed these lists as they appeared as of February 15, 2007;
additional breaches that occurred during the 2-year period we reviewed may
have been subsequently added as they were discovered. Our analysis
eliminated overlap among the three lists; the 572 breaches we cite
represent unique breaches that appeared on at least one list.

  Federal Law Enforcement Agencies

Officials at federal law enforcement agencies told us that each year they
conduct a significant number of criminal investigations that involve
alleged breaches of sensitive personal information. For example, officials
of the FBI's Cyber Division told us that presently it has more than 1,300
pending cases of computer or network intrusions where data breaches
resulted from unauthorized electronic access to computer systems, such as
hackings, at public and private organizations.19 Officials at the Secret
Service, which investigates certain cases where financial information has
been lost or stolen, told us that in 2006, the service opened 327 cases
involving network intrusions or other breaches at retailers, banks, credit
card processors, telephone companies, educational institutions, and other
organizations. Officials noted that they have seen a steady increase in
the number of data breaches since 1986, when they began tracking computer
fraud violations. Investigators at USPIS, the division of the U.S. Postal
Service that investigates mail fraud, external mail theft, fraudulent
changes of addresses, and other postal-related crimes, told us that the
agency does not specifically track the number of data breaches in the
private sector. However, despite limited data, investigators said their
impression is that such data breaches likely occur frequently.

18There were 83 million records collectively reported to have been
affected by the 572 breaches. However, in some cases, the number of
records affected was unknown or unreported, and the total does not reflect
those breaches. Also, the number of breached records containing personal
information may not be the same as the number of individuals affected by
breaches because some individuals may be victims of more than one breach
or may have multiple records compromised in a single breach. Finally, in
addition to the 83 million records, as many as 40 million additional
records may have been affected by a single breach involving the credit
card processor CardSystems, although the exact number of affected records
is unclear. In a complaint following the breach, FTC alleged that a hacker
obtained unauthorized access to magnetic stripe data for tens of millions
of credit and debit cards. However, according to testimony by a
CardSystems official, only 263,000 of these records (containing 239,000
discrete account numbers) included sensitive personal information.

19According to these officials, not all 1,300 pending computer intrusion
cases necessarily involved breaches that compromised sensitive personal
information, although the vast majority have. The term hacking is commonly
used to refer to accessing a computer system without authorization, with
the intention of destroying, disrupting, or carrying out illegal
activities on the network or computer system.

  House Government Reform Committee and DHS

To obtain information on the prevalence of data breaches at federal
agencies, in July 2006 the House Government Reform Committee asked federal
agencies to provide details about incidents involving the loss or
compromise of any sensitive personal information held by an agency or
contractor from January 1, 2003, through July 10, 2006. Our analysis of
the committee's report found that 17 agencies reported that they
experienced at least one breach and, collectively, the agencies reported
to the committee more than 788 separate incidents.20

The Federal Information Security Management Act of 2002 requires all
federal agencies to report computer security incidents to a federal
incident response center.21 The U.S. Computer Emergency Readiness Team--a
component of DHS that monitors computer security incidents at federal
agencies--serves as this response center. As such, data breaches at
federal agencies involving certain sensitive information must be reported
to the team within 1 hour of discovery of the incident.22 DHS staff told
us that they receive information about breaches at federal agencies on a
daily basis. In fiscal year 2006, the center tracked 477 incidents at 59
federal agencies or at federal contractors with access to government-owned
data, according to information available as of January 29, 2007. In
addition, a March 2007 audit investigation found that at least 490 laptop
computers owned by the Internal Revenue Service and containing taxpayer
information had been lost or stolen since 2003.23

20Government Reform Committee, U.S. House of Representatives, Staff
Report: Agency Data Breaches Since January 1, 2003 (Washington, D.C.: Oct.
13, 2006). The federal agencies covered in the report were the Departments
of Agriculture, Commerce, Defense, Education, Energy, Health and Human
Services, Homeland Security, Housing and Urban Development, Interior,
Justice, Labor, State, Transportation, Treasury, and Veterans Affairs, as
well as the Office of Personnel Management and the Social Security
Administration. In addition to 788 incidents reported by 16 federal
agencies, the Committee received information on data breaches from the
Department of Veterans Affairs, which the report characterized only as
"hundreds" of incidents.

21Pub. L. No. 107-347, tit. III, 116 Stat. 2946 (Dec. 17, 2002), codified
at 44 U.S.C. S 3541-3549; 40 U.S.C. S 11331.

  Federal Banking Regulators

The 2005 guidance issued by the five federal banking regulators provided
that a depository institution should notify its primary federal regulator
when it becomes aware of an incident involving unauthorized access to or
use of sensitive customer information.24 The guidance applies to breaches
that have occurred at the financial institutions themselves, as well as
third-party entities such as data processors that act as service providers
and maintain customer information.25 The five regulators differ in their
methods and criteria for tracking breaches, but collectively they have
tracked several hundred breaches over the past few years at roughly 17,000
institutions they supervise and at third-party entities.26 For example,
the Federal Deposit Insurance Corporation (FDIC)--the primary federal
supervisor for more than 5,000 state-chartered banks that are not members
of the Federal Reserve System--received reports of 194 breaches at its
regulated institutions from May 2005 through December 2006, as well as
reports of 14 breaches at third-party companies that also affected these
institutions' customers. Similarly, officials at the Office of Thrift
Supervision--which supervises more than 860 savings associations--told us
that from April 2005 through December 2006, 56 of its institutions
reported breaches at the institution itself and approximately 72 reported
breaches at third-party entities that maintained their customer
information.

22Office of Management and Budget Memorandum for Chief Information
Officers, Reporting Incidents Involving Personally Identifiable
Information and Incorporating the Cost for Security in Agency Information
Technology Investments, M-06-19 (July 12, 2006). The U.S. Computer
Emergency Readiness Team defines a computer security incident as "a
violation or imminent threat of violation of computer security policies,
acceptable use policies, or standard computer security practice" and the
Office of Management and Budget requires reporting if the incident
includes personally identifiable information, which under its definition
refers to "any information about an individual maintained by an agency,
including, but not limited to, education, financial transactions, medical
history, and criminal or employment history and information which can be
used to distinguish or trace an individual's identity, such as their name,
Social Security number, date and place of birth, mother's maiden name,
biometric records, etc., including any other personal information which is
linked or linkable to an individual."

23Treasury Inspector General for Tax Administration, The Internal Revenue
Service Is Not Adequately Protecting Taxpayer Data on Laptop Computers and
Other Portable Electronic Media Devices, Ref. No. 2007-20-048 (Washington,
D.C.: Mar. 23, 2007).

2470 Fed. Reg. 15736 (Mar. 29, 2005) and 70 Fed. Reg. 22764 (May 2, 2005).

25Only data breaches at the financial institutions and at third-party
entities that are their service providers and maintain their customer
information are subject to the guidance; this requirement is codified at
12 C.F.R. Pt. 30, App. B, Supp. A S II(A)(2); 12 C.F.R. Pt. 208, App. D-2,
Supp. A S II(A)(2); 12 C.F.R. Pt. 225, App. F, Supp. A S II(A)(2); 12
C.F.R. Pt. 364, App. B, Supp. A S II(A)(2); 12 C.F.R. Pt. 570, App. B,
Supp. A S II(A)(2); and 12 C.F.R. Pt. 748, App. B S II(A)(2). However,
data collected by the regulators may also include some breaches that
affected their institutions but were not covered by the guidance.

  State Agencies

Some states require entities experiencing data breaches to report them to
designated state agencies.27 For example, the New York State Information
Security Breach and Notification Act requires entities that experience
security breaches to notify the state Attorney General's Office, Consumer
Protection Board, and the state Office of Cyber Security and Critical
Infrastructure Coordination in cases when New York residents must be
notified.28 Such data breaches include the unauthorized acquisition of
computerized data that compromises the security, confidentiality, or
integrity of unencrypted private information. Officials of the Office of
New York State Attorney General told us that from December 7, 2005,
through October 5, 2006, their office received notice of 225 breaches.
Similarly, a North Carolina law requires that breaches of personal
information (maintained in computerized, paper, or other media) affecting
at least 1,000 persons be reported to the Consumer Protection Division of
the state Office of the Attorney General.29 An official in that office
told us that from December 2005 through December 2006, it had received
reports of 91 breach incidents.

26Regulators note that while they track breaches occurring at third-party
service providers involving customer information of regulated financial
institutions, these breaches are typically due to lapses in data security
by the third-party entity and not the financial institution itself.

27We did not determine the precise number of states with centralized
reporting requirements. For illustrative purposes, we obtained information
on data breaches from New York and North Carolina because they are two
large states known to require that data breaches be reported to state
agencies.

28N.Y. Gen. Bus. Law S 899-aa.

29N.C. Gen. Stat S 75-65.

  Other Sources

Information that we obtained from several other sources suggests that
breaches of sensitive personal information occur with some frequency
across a variety of sectors. For example,

           o EDUCAUSE, a nonprofit association that addresses technology
           issues in higher education, conducted a survey in 2005 on data
           security at higher education institutions in the United States and
           Canada. Twenty-six percent of the 490 institutions that responded
           said they had experienced a security incident in the past year
           that resulted in the compromise of confidential information.30

           o The American Hospital Association collected information, at our
           request, in October 2006 from a nonrepresentative group of 46
           large hospitals on breaches of sensitive personal information
           (excluding medical records) that they had experienced since
           January 2003. Collectively, 13 of the 46 hospitals reported a
           total of 17 data breach incidents.31

           o The Ponemon Institute, a private company that researches privacy
           and security practices, conducted a survey of 51,433 U.S. adults
           and received responses from 9,154 (a response rate of about 18
           percent). About 12 percent of the survey respondents said they
           recalled receiving notification of a data security breach
           involving their personal information.32

           o The CMO Council, an organization serving marketing executives,
           reported that 16 percent of consumers who responded to a Web-based
           panel reported that a company had lost or compromised their
           personal, financial, or medical information. An additional 32
           percent of respondents said they were not sure.33
			  
30EDUCAUSE Center for Applied Research, Safeguarding the Tower: IT
Security in Higher Education 2006, Volume 6, 2006.

31The association received information from 46 of the 78 hospitals it
surveyed, a response rate of 59 percent. As agreed in advance, to preserve
confidentiality the association provided us with a summary of their
findings but did not identify the hospitals, and we did not independently
verify the data.

32Ponemon Institute, LLC, National Survey on Data Security Breach
Notification (Sept. 26, 2005). The reliability of this study's findings
may be limited by a low survey response rate.

           o Several other studies, while not focusing specifically on
           breaches of sensitive personal information, have found more
           generally that information security vulnerabilities are widespread
           among U.S. and global companies.34

Information from multiple sources indicates that data breaches at
companies, government agencies, retailers, and other entities have
occurred frequently in recent years, involving millions of records of
sensitive personal information. We have reported in the past that no
federal law explicitly requires most companies and other entities to
safeguard all of the sensitive personal information they may hold. We also
have suggested that to ensure that sensitive personal information is
protected on a more consistent basis, Congress should consider expanding
requirements to safeguard such information.35 The frequency of data
breaches identified in this report underscores the need for entities in
the public and private sectors to improve the security of sensitive
personal information and further corroborates that additional federal
action may be needed in this area.

Source, Cause, Size, and Content of Breaches Have Varied Widely

According to government officials, researchers, and media reports, data
breaches have occurred among a wide variety of entities and as a result of
both intentional actions and accidental losses. These breaches also have
varied in size and in the types of data compromised.

  Type of Entity

Data breaches have been reported at a wide range of public and private
institutions, including federal, state, and local government agencies;
public and private colleges and universities; hospitals and other medical
facilities; retailers; banks and other financial institutions; information
resellers; and others. For example, in the weeks leading up to the highly
publicized 2005 CardSystems breach, the media also had reported breaches
at, among other entities, a large hospital, a university, a global
financial institution, a federal regulatory agency, and a major technology
company.

33CMO Council, Securing the Trust of Your Brand: How Security and IT
Integrity Influence Corporate Reputation, September 2006. The reliability
of this study's findings may be limited because they are based on a
self-selected group of respondents to a Web-based panel. Also, we were
unable to determine a response rate because, according to a CMO Council
representative, the total number of survey respondents was not available.

34For example, see Deloitte, 2006 Global Security Survey (2006); Small
Business Technology Institute, Small Business Information Security
Readiness (San Jose, California: July 2005); Ponemon Institute, LLC, U.S.
Survey: Confidential Data at Risk (Aug. 15, 2006); and Ponemon Institute,
LLC, Benchmark Study of European and U.S. Corporate Privacy Practices
(Apr. 26, 2006).

35GAO, Personal Information: Key Federal Privacy Laws Do Not Require
Information Resellers to Safeguard All Sensitive Data, [39]GAO-06-674
(Washington, D.C.: Jun. 26, 2006).

According to Attrition, of the breaches it tracked as reported in the news
media in 2005 and 2006, 33 percent of the breaches occurred at educational
institutions, 32 percent at financial services institutions, 25 percent at
government agencies, and 10 percent at medical facilities, although
breaches reported in the news media may not be representative of all
breaches.36 Similarly, the data security firm ID Analytics examined 70
data breaches that were reported by the news media from February through
September 2005. According to company officials, 46 percent of these
breaches occurred at educational institutions, 16 percent at financial
institutions, 14 percent at retailers, 11 percent at government agencies,
7 percent at medical facilities, and 6 percent at information resellers.37

Another way to analyze where data breaches have occurred is to look at the
number of records breached (as opposed to the number of breaches
themselves). Our analysis of the list maintained by Attrition found that
54 percent of breached records involved financial institutions, 34 percent
involved government agencies, 4 percent involved educational institutions,
and 3 percent involved medical facilities. ID Analytics' report found that
57 percent of breached records involved financial institutions, 22 percent
involved retailers, 13 percent involved educational institutions, 4
percent involved information resellers, 2 percent involved government
agencies, and 2 percent involved medical facilities.

  Cause of Breach

According to government officials, researchers, and media reports, data
breaches of sensitive personal information have occurred as a result of
both intentional actions as well as negligence or accidental losses. In
some cases, individuals intentionally steal information for the purpose of
committing fraud or identity theft. Breaches involving intentional actions
have included:

36For our analysis, we used the categories provided by Attrition for the
industry sector where the breach occurred. We did not independently verify
the accuracy of these categorizations.

37ID Analytics, Inc., National Data Breach Analysis (San Diego,
California: January 2006). The data we cite reflect a combination of data
presented in the report and additional data provided to us by ID
Analytics.

           o Hacking, or accessing computer systems without authorization.
           For example, in 2007 the retailer TJX Companies reported
           unauthorized intrusions into its computer systems that may have
           breached millions of customers' credit card and driver's license
           information.

           o Employee theft. For example, in 2006, a former employee of the
           American Red Cross pled guilty to stealing personally identifiable
           information from a blood donor database.

           o Theft of physical equipment. In 2005, for instance, a laptop
           containing the names and SSNs of more than 98,000 students,
           alumni, and others was stolen from the University of California at
           Berkeley.

           o Deception or misrepresentation to obtain unauthorized data. In
           2005, the information reseller ChoicePoint acknowledged that the
           personal records it held on approximately 162,000 consumers had
           been compromised by individuals who posed as legitimate
           subscribers to the company's information services.

Breaches involving negligence or accidental losses of data have included
the following:

           o Loss of laptop computers or other hardware. For example, in
           2006, the Department of Labor reported that an employee lost a
           laptop containing personal information on 1,137 individuals.

           o Loss of data tapes. For example, in 2004, Bank of America lost
           backup tapes containing personal information of 1.2 million
           government charge card holders while the tapes were being
           transported to a data center.

           o Unintentional exposure on the Internet. In 2006, according to
           media reports, the U.S. Department of Education left unprotected
           on a Web site the personally identifiable information, including
           SSNs, of up to 21,000 recipients of federal student loans.

           o Improper disposal of data, such as leaving sensitive personal
           data on unshredded documents in a publicly accessible dumpster.

We did not identify comprehensive data that reliably provide overall
statistics on the causes of known data breaches. However, our review of
the 24 largest data breaches reported in the news media (discussed in more
detail later in this report) found that 12 breaches apparently involved
intentional acts by hackers or employees illegally accessing or using
data, 5 involved stolen laptops or other computer equipment, 4 involved
lost computer backup tapes, 2 involved the use of deception to gain access
to data, and 1 involved the possible unauthorized disclosure of data. In
addition, some studies indicate that most breaches reported in the news
media resulted from intentional acts rather than accidental occurrences
such as a lost laptop computer. For example, in its study of 70 breaches,
ID Analytics determined that 48 involved thefts committed with the
apparent intention of accessing sensitive data. Eleven of the breaches
involved thefts where sensitive consumer information was apparently stolen
inadvertently as part of another crime (such as the theft of a laptop
computer for its resale value), and another 11 breaches involved
accidental loss (such as misplacement of a laptop computer). However,
these data may overrepresent the proportion of all breaches that involve
criminal activity, as such breaches are probably more likely than
accidental losses to be reported to authorities and by the news media.

  Number of Records Breached

Our analysis of the list maintained by Attrition of breaches reported by
the news media found the median number of records breached to be 8,650.
However, these data breaches varied considerably in size--ranging, for
example, from a breach involving 10 records at a law firm to a breach
involving as many as tens of millions of records at a credit card
processing company. The breaches involving federal agencies that were
reported to the House Government Reform Committee also varied in size--for
example, several affected fewer than five records, while a breach at VA
affected 26.5 million records.

  Types of Data Breached

Comprehensive information does not exist on the types of data involved in
all known data breaches. Among the list maintained by Attrition of
breaches reported by the news media in 2005 and 2006--which may not be
representative of all breaches--more than half involved SSNs and 11
percent involved credit card numbers (and 3 percent of the total involved
both). In the remaining breaches, other types of account or personal
information were involved, or the type of data breached was not reported.
Logically, there may be an association between the type of data
compromised and the type of entity experiencing the breach. For example,
several educational institutions have experienced breaches of SSNs, which
they may maintain as student identifiers, and several retail stores have
experienced breaches of credit card numbers, which they often maintain on
their customers.

Consequences of Data Breaches Are Not Fully Known, but Clear Evidence of
Identity Theft Has Been Found in Relatively Few Breaches

Comprehensive information on the outcomes of data breaches is not
available. Several cases have been identified in which a data breach
appears to have resulted in identity theft, but available data and
information from law enforcement and industry association representatives
indicated that most breaches have not resulted in detected incidents of
identity theft. For example, of 24 very large breaches we reviewed, 3
appeared to have resulted in fraud on existing accounts and 1 in the
unauthorized creation of new accounts. Determining the link between data
breaches and identity theft is challenging because, among other things,
identity theft victims often do not know how their personal information
was obtained. However, the circumstances of a breach, including the type
of information compromised and how the breach occurred, can greatly affect
the potential risk of identity theft.

Federal Law Enforcement Agencies and Industry Associations Identified Limited
Instances of Breaches Leading to Identity Theft

In general, representatives of law enforcement agencies, industry and
trade associations, and consumer and privacy advocacy organizations told
us that no comprehensive data are available on the consequences of data
breaches. Several cases have been identified where there is evidence that
a data breach resulted in identity theft, including account fraud or
unauthorized creation of new accounts. At the same time, available data
and information from the officials we contacted indicated that most
breaches have not resulted in detected incidents of identity theft.

We asked representatives of the FBI, Secret Service, USPIS, and
Immigration and Customs Enforcement--a component of DHS that has
investigated cases where stolen identities were used to secure jobs--the
extent to which data breaches they investigated resulted in some form of
identity theft. Representatives of all of these agencies told us that
their investigations of data breaches do not typically allow them to fully
ascertain how stolen data are used. Similarly, they noted that
investigations of identity theft do not always reveal the source of the
data used to commit the crime.

However, the representatives were able to provide us with a limited number
of examples in which data breaches they investigated had allegedly
resulted in some form of identity theft. For example, in a 2006
investigation by USPIS, an employee of a credit card call center allegedly
compromised at least 35 customers' accounts and used some of the
information to purchase approximately $65,000 in gift cards. The
representatives of federal law enforcement agencies noted that cases in
which data breaches have been linked to identity theft often have involved
instances of unauthorized access by employees. For example, an official at
Immigration and Customs Enforcement stated that her agency, in cooperation
with other agencies, has investigated cases in which government employees
allegedly had improperly accessed and sold sensitive personal information
that was then used by illegal immigrants to secure employment.

In addition, in 2005 FTC settled charges with BJ's Wholesale Club in which
alleged security breaches resulted in several million dollars in
fraudulent purchases using customers' credit and debit card data.38 As
discussed later in this report, FTC has also taken enforcement actions
related to data breaches at several other companies, including
ChoicePoint, CardSystems, and DSW, in which it uncovered evidence that the
breaches resulted in identity theft.

Many of the law enforcement officials said that, based on their
experience, data breaches that result in harm have usually involved fraud
on existing accounts (such as credit card fraud) rather than the
unauthorized creation of new accounts. Secret Service representatives
noted that using illicit credit and debit card numbers and bank account
information is much easier and less labor intensive than using personally
identifiable information to fraudulently open new accounts. Officials at
Secret Service, FBI, and USPIS all said that identity theft involving the
creation of new accounts often results not from data breaches, but from
other sources, such as retrieving personal information by sifting through
a family's household trash.

In examining a selection of five breaches that occurred from 2003 through
2005 that were reported as having involved five federal
agencies--Department of Justice, FDIC, Internal Revenue Service, National
Park Service, and the Navy--we found that the circumstances behind these
breaches varied widely. At least two of the breaches occurred at vendors
or contractors that held sensitive data on agency employees, rather than
at the agency itself. In addition, we found that a breach reported in the
news media as having involved the National Park Service actually involved
a not-for-profit organization that manages eParks, according to a
representative of that organization. Four of the five breaches reported as
having involved federal agencies were not believed to have resulted in
identity theft, according to officials of the entities involved. The
breach at FDIC resulted in an estimated 27 cases of identity theft when
data inappropriately accessed by a former FDIC intern were used to take
out more than $425,000 in fraudulent loans in the names of FDIC employees,
according to agency officials.39

38In the Matter of BJ's Wholesale Club, Inc., F.T.C. No. 0423160 (2005). A
consent agreement does not constitute an admission of a violation of law.

Industry and trade associations representing entities that maintain large
amounts of information--banks, retailers, colleges, information resellers,
and hospitals--told us that they had limited knowledge about the harm
caused by data breaches that occur in their industries. However, in some
cases, they provided information or anecdotal evidence on the extent to
which such breaches may have led to some form of identity theft. For
example, the 46 hospitals that the American Hospital Association surveyed
at our request reported that of 17 breaches that had occurred since 2003,
three had resulted in fraudulent activity on existing accounts and another
three resulted in other forms of identity theft, including one case where
the information was used to file false income tax refunds. The identity
theft in these cases involved small numbers of victims--usually just one.

Representatives of the American Council on Education and two other higher
education associations stated that while data breaches at colleges and
universities were not uncommon, they were aware of little to no identity
theft that had resulted from such breaches. Representatives of the
American Bankers Association, the National Retail Federation, and the
Consumer Data Industry Association told us they were unable to determine
how prevalent data breaches are among their institutions or how often such
breaches lead to consumer harm. Representatives at the National Retail
Federation noted that breaches at retailers may be more likely to result
in fraud on existing accounts than in new account creation, since most
retailers do not maintain the personally identifiable information needed
to steal someone's identity.

39According to an FDIC representative, the agency took several steps to
address the possible misuse of employee information, including promptly
notifying affected employees and offering them 2 years of credit
monitoring services.

Of 24 Large Publicly Reported Breaches, 4 Apparently Resulted in Known Cases of
Identity Theft

Using lists of data breaches compiled by the Identity Theft Resource
Center, Privacy Rights Clearinghouse, and Congressional Research Service,
we identified the 24 largest breaches (measured by number of records) that
were reported in the news media from January 2000 through June 2005.40 To
gather information on these incidents, we interviewed or collected written
responses from representatives of the entity experiencing the breach and
reviewed publicly available information, such as media reports, news
releases, testimonies, and court documents. In some cases, when feasible,
we also spoke with law enforcement investigators. We identified those
cases where this information collectively indicated that the breach
appeared to have resulted in some form of identity theft. Ultimately, the
determination of whether particular conduct violated a law prohibiting
identity theft would be a matter of law for the courts.

Although these lists characterized each of these 24 incidents as data
breaches, the circumstances of the incidents varied. While 19 of the
incidents clearly met our definition of data breach (i.e. unauthorized or
unintentional exposure, disclosure, or loss of sensitive personal
information), four cases involved hackers who may or may not have actually
accessed sensitive information. In one other incident, a university
employee with access to sensitive personal data was indicted on unrelated
fraud charges. A university official told us he did not believe this
incident should necessarily be characterized as a data breach since there
was no evidence the employee actually misused university data.

The available evidence that we reviewed indicated that 18 of these 24
breaches were not known to have resulted in any identity theft. As shown
in table 1, three breaches were believed to have resulted in account fraud
and one resulted in the unauthorized creation of new accounts. In two
other cases, we were not able to gather sufficient information on whether
harm appeared to have resulted from the breach. Further, because of the
challenges in linking data breaches with identity theft, in some cases our
review may not have uncovered instances of harm potentially resulting from
these breaches. In some instances, investigators or company
representatives reported that they were able to determine with a high
degree of certainty--through forensic investigation or other means--that
unauthorized parties had not accessed the data. In other instances, these
representatives said that they were not aware of any account fraud that
resulted, but they acknowledged that there was no way to know for sure.
Moreover, determining potential harm may be particularly challenging with
very large breaches because the volume of records involved can make it
difficult to link individual victims to the breach.

40These three organizations periodically update their lists by adding
breaches they learn about that occurred in the past, including some that
occurred between January 2000 and June 2005. Our list of the 24 largest
media-reported breaches was based on information provided by these lists
as of August 2006. We were not aware of the Attrition list at the time we
made our selection. See Congressional Research Service, Personal Data
Security Breaches: Context and Incident Summaries, Order Code RL33199
(Washington, D.C.: Dec. 16, 2005). Because our time frame covered only
breaches that occurred on or before June 30, 2005, our list does not
include highly publicized breaches that occurred subsequently, such as
those involving the Department of Veterans Affairs and the TJX Companies.
Several banks have reported fraudulent transactions on existing accounts
resulting from the TJX breach, according to a January 24, 2007, press
release by the Massachusetts Bankers Association.

Table 1: Twenty-Four Large Publicly Reported Data Breaches and Evidence of
Resulting Identity Theft, January 2000 - June 2005

The fact that we did not identify evidence of identity theft from a breach
does not necessarily mean that no such harm has occurred or will occur in
the future.
                                                            Available         
                                                            evidence          
                                                                              
            Type of                                         of identity       
Yeara    organization       Nature of breach             theft?b           
2000     Retail             Hacking                      Account fraud     
2000     Retail             Hacking                      None identified   
2002     Healthcare         Stolen computer equipment    None identified   
2003     Higher education   Stolen computer equipment    None identified   
2004     Financial services Stolen computer equipment    None identified   
2004     Higher education   Hacking                      None identified   
2004     Higher education   Hacking                      None identified   
2004     Higher education   Hacking                      None identified   
2004     Financial services Lost data tapes              None identified   
2005     Financial services Hacking                      Account fraud     
2005     State government   Hacking                      None identified   
2005     Information        Deception/Misrepresentation  Unauthorized new  
            services                                        accounts          
2005     Higher education   Hacking                      None identified   
2005     Higher education   Stolen computer equipment    None identified   
2005     Retail             Hacking                      Account fraud     
2005     Information        Deception/Misrepresentation  Unknown           
            services                                                          
2005     Healthcare         Stolen computer equipment    None identified   
2005     Retail             Hacking                      Unknown           
2005     Financial services Lost data tapes              None identified   
2005     Financial services Employee crime               None identified   
2005     State government   Hacking                      None identified   
2005     Media              Lost data tapes              None identified   
2005     Financial services Lost data tapes              None identified   
2005     Higher education   Otherc                       None identified   

Source: GAO.

Note: To identify the 24 largest data breaches reported in the news media
from January 2000 through June 2005, GAO analyzed lists of such breaches
maintained by Identity Theft Resource Center, Privacy Rights
Clearinghouse, and Congressional Research Service.

aYear breach occurred or was publicized.

bThe presence or lack of evidence of identity theft resulting from a
breach was based on our review of news reports and other publicly
available information, as well as interviews, as feasible, with
representatives of entities experiencing the breach and law enforcement
officials investigating the breach. The fact that we were unable to
identify evidence at this time of identity theft resulting from a breach
does not mean that no such harm has occurred or that none will occur in
the future. Further, factual determinations of the existence and cause of
identity theft in any particular case are matters for the courts to
decide.

cIn this case, a former university employee with access to sensitive
personal information had been indicted on bank fraud charges unrelated to
the university. Some press reports characterized this as a breach, but
according to a representative of the university, there is no evidence that
the employee misused university data.

The one large breach we identified that apparently resulted in the
unauthorized creation of new accounts involved ChoicePoint, an information
reseller. In 2005, the company acknowledged that the personal records it
held on approximately 162,000 consumers had been compromised by
individuals who posed as legitimate subscribers to the company's
information services. FTC reached a civil settlement in 2006 with the
company that established a fund for consumer redress to reimburse
potential victims of identity theft, and the agency has worked with law
enforcement officials to identify such victims.41

The three large breaches we identified that appeared to result in fraud on
existing accounts included the following:

           o CardSystems, a credit card payment processor, reported a May
           2005 breach in which a hacker accessed data such as names, card
           account numbers, and expiration dates. The total number of
           compromised accounts is unclear. FTC staff alleged in a 2006 civil
           complaint that the breach had compromised data associated with
           tens of millions of credit and debit cards, but a CardSystems
           official stated in congressional testimony that only 239,000
           accounts were compromised. Officials of the Office of the
           Comptroller of the Currency--who surveyed the national banks they
           supervise in order to determine the amount of fraudulent charges
           that resulted from the breach--said that customers of 110 banks
           were affected by this incident and losses of more than $13 million
           in fraudulent charges on customers' cards were reported by 24 of
           these institutions.
			  
41United States v. ChoicePoint, Inc., No. 1:06-cv-00198-JTC (N.D. Ga.,
Feb. 15, 2006). As part of the settlement, ChoicePoint admitted no
violations of the law. According to ChoicePoint, the company has
subsequently taken steps to enhance its customer screening process and to
assist affected consumers. FTC staff told us that law enforcement
officials have determined that as many as 2,900 people have experienced
the fraudulent creation of new accounts as a result of the breach.
According to a ChoicePoint official, the criminal indictments indicated
that 46 people may have been defrauded, but the accused individuals may
not have used data acquired from ChoicePoint in all the crimes cited in
the indictments.

           o DSW, a shoe retailer, said in an April 2005 news release that it
           had experienced a data breach in which a hacker accessed the names
           and card numbers associated with 1.4 million credit and debit card
           transactions at 108 of its stores, as well as checking account
           numbers and driver's license numbers from 96,000 check
           transactions. According to a complaint filed by FTC in March of
           2006, there allegedly have been fraudulent transactions on some of
           these accounts.

           o CD Universe, an Internet-based music store, reportedly
           experienced a breach in December 1999 in which a hacker accessed
           as many as 300,000 names, addresses, and credit card numbers from
           the company Web site, according to media reports and a company
           official. The hacker allegedly used some of the stolen credit card
           numbers to obtain money for himself.42

Challenges Exist in Determining the Link between Data Breaches and Identity
Theft

Determining the link between data breaches and identity theft is
challenging for several reasons. First, identity theft victims often do
not know how their personal information was obtained. According to FTC, in
approximately 65 percent of the identity theft complaints it received from
October 1, 2005, through August 31, 2006, the victim did not know or
report how the information was compromised. Second, victims may
misattribute how their data were obtained. For example, federal officials
and representatives of a private group that assists victims said that
consumers who are notified of a breach often assume that any perceived
mistakes on their credit card statements or credit report were a result of
the breach. As a result, no government agency maintains comprehensive data
on the underlying cause of identity theft. FTC told us that its Identity
Theft Data Clearinghouse is limited to self-reported complaints and
therefore does not contain statistically reliable information that would
allow the agency to determine a link between data breaches and identity
theft. Similarly, according to FBI, data maintained by the Internet Crime
Complaint Center does not include information sufficient to determine the
link between data breaches and identity theft.

42This breach occurred in December 1999 but was included in the 24
breaches we reviewed because it was reported in the media in January 2000.

Third, law enforcement officials told us that in some cases, stolen data
may be held for up to a year or more before being used to commit identity
theft. Further, once stolen data have been sold or posted on the Web,
fraudulent use of that information may continue for years. As a result,
studies that attempt to measure the harm resulting from data breaches
cannot necessarily rule out all future harm. Finally, conducting
comprehensive studies of data breaches and identity theft can be hindered
by issues of privacy and confidentiality. For example, companies that have
experienced breaches may be unable or unwilling to provide information
about affected individuals to researchers.

Some studies conducted by private researchers have sought to determine the
extent to which data breaches result in identity theft, but our review
found them to contain methodological limitations.43 One research firm
conducted a study of four data breaches, analyzing credit and other
application data for suspicious relationships that indicated fraud.44 The
study estimated that no more than 0.10 percent of individuals whose data
had been breached experienced resulting identity theft in the form of
unauthorized new account creation. However, because the study reviewed
only four data breaches, it cannot be considered representative of other
breaches. Moreover, two of these breaches did not involve personally
identifiable information and thus would not be expected to create a risk
of fraud involving new account creation.

Another private research firm surveyed approximately 9,000 individuals
about whether they had ever received a notification from an organization
about the loss or theft of their personal information.45 Of the
approximately 12 percent of individuals who reported they had received
such a notification, 3 percent--or 33 people--said they believed they had
suffered identity theft as a result. However, these data are subject to
limitations; among other things, individuals are often unaware of whether
any fraud they have suffered was, in fact, due to a data breach. A third
firm projected in a study that 0.8 percent of consumers whose information
a data breach compromised would experience fraud as a result.46 However,
we question the reliability of this estimate, in part because of
assumptions made about the number of consumers affected by data breaches.

43Although we found limitations in how these studies linked data breaches
and identity theft, we determined other aspects of these studies to be
sufficiently reliable, and we refer to them elsewhere in this report.

44ID Analytics, Inc., National Data Breach Analysis (2006).

45Ponemon Institute, National Survey (2005). As noted earlier, this study
may also be limited by a low survey response rate.

Type of Data Compromised and Other Factors Influence Potential for Resulting
Consumer Harm

The type of data compromised in a breach can effectively determine the
potential harm that can result. For example, credit or debit card
information such as card numbers and expiration dates generally cannot be
used alone to open unauthorized new accounts. Some of the largest and most
highly publicized data breaches in recent years largely involved credit or
debit card data rather than personally identifiable information. As a
result, these breaches put affected consumers at risk of account fraud but
not necessarily at risk of fraud involving unauthorized creation of new
accounts--the type of identity theft generally considered to have a more
harmful direct effect on consumers. While credit and debit card fraud is a
significant problem--the FTC estimates it results in billions of dollars
in losses annually--existing laws limit consumer liability for such fraud
and, as a matter of policy, some credit and debit card issuers may
voluntarily cover all fraudulent charges.47 In contrast, the unauthorized
creation of new accounts--such as using someone else's identity to open
credit card or bank accounts, originate home mortgages, file tax returns,
or apply for government benefits--can result in substantial financial
costs and other hardships.

In addition to the type of data compromised in a breach, several
additional factors can influence the extent to which a breach presents the
risk of identity theft. These include the following:

           o Intent. Breaches that are the result of intentional acts--such
           as hacking into a server to obtain sensitive data--generally are
           considered to pose more risk than accidental breaches such as a
           lost laptop or the unintentional exposure of sensitive data on the
           Internet, according to federal agency officials. However, in some
           cases, such as the theft of a laptop containing personal
           information, it may be unknown whether the laptop was stolen for
           the hardware, the personal data, or both.

           o Encryption. Encryption--encoding data so that it can only be
           read by authorized individuals--can in some cases prevent
           unauthorized access. However, some forms of encryption are more
           effective than others, and encryption does not necessarily
           preclude fraudulent use of data--for example, if the key used to
           unencrypt the data is also compromised.

           o Hardware requirements. Data that only can be accessed using
           specialized equipment and software may be less likely to be
           misused in the case of a breach. For example, some entities that
           have lost data tapes have stated that criminals would require
           specific data reading equipment and expertise in how to use it to
           access the information.

           o Number of records. Larger breaches may pose a greater overall
           risk that at least one individual would become a victim of
           identity theft. At the same time, given the resources needed to
           commit identity theft, breaches of very large numbers of records
           may pose less risk to any one individual whose data were
           compromised.

46Javelin Strategy & Research, Data Breaches and Identity Fraud:
Misunderstanding Could Fail Consumers and Burden Businesses (Pleasanton,
California, August 2006).

47For unauthorized credit card charges, consumer liability is limited to a
maximum of $50 per account, 15 U.S.C. S 1643. For unauthorized ATM or
debit card transactions, the Electronic Fund Transfer Act limits consumer
liability, depending on how quickly the consumer reports the loss or theft
of the card. Pub. L. No. 90-321, tit. IX, as added Pub. L. No. 95-630,
tit. XX, S 2001, 92 Stat. 3728 (Nov. 10, 1978); 15 U.S.C. S 1693g.
Consumers may incur additional costs if they inadvertently pay charges
they did not incur. In addition, account fraud can cause inconvenience or
temporary hardship--such as losing temporary access to account funds or
requiring the cancellation and reactivation of cards and the redirecting
of automatic payments and deposits.

Breach Notification Requirements Can Serve to Encourage Better Data Security
Practices and Alert Consumers, but They Also Present Costs and Challenges

Breach notification requirements have several potential benefits,
including creating incentives for entities to improve their data security
practices (and thus prevent potential breaches from occurring), allowing
affected consumers to take measures to prevent or mitigate identity theft,
and serving to respect individuals' basic right to know when their
personal information is compromised. At the same time, breach notification
requirements present costs, both for developing compliance strategies and
for actual notifications in the event of a breach. Further, there is the
risk of overnotification, or inundating consumers with frequent
notifications of breaches that may present little or no risk of identity
theft or other harm. Thus, policymakers face the challenge of setting a
notification standard that allows individuals to take steps to protect
themselves where the risk of harm exists, while ensuring they are only
notified in cases where the level of risk warrants such action.

Notification Requirements May Create Incentives for Improved Data Security and
Allow Consumers the Opportunity to Mitigate Risks

According to our review of studies and interviews with representatives in
government, academia, and private industry, breach notification
requirements have several potential benefits, as follows:

           o Incentives for Improved Data Security. Breach notification
           requirements can provide an incentive for companies and other
           entities to increase their data security measures to avoid the
           possible financial and reputational risks that can be associated
           with a publicly reported data breach.48 Representatives we
           contacted in the private, nonprofit, and government sectors told
           us that they believe that existing breach notification
           requirements in state laws, or the breach notification provisions
           in federal banking regulatory guidance, have provided entities
           with incentives to improve data security practices. For example,
           some representatives of companies and other organizations noted
           that passage of state notification laws led to companies
           reexamining data security procedures and making improvements, such
           as encrypting sensitive data and restricting consumer data that
           can be accessed online. Similarly, federal banking regulators told
           us that they believe their notification guidance has motivated
           regulated institutions to enhance data security. For example,
           according to officials at the Office of Thrift Supervision, its
           institutions have taken steps such as improving electronic
           firewalls and implementing formal incident response reporting
           systems.

           o Prevention of Identity Theft. Breach notification can provide
           consumers with the opportunity to take steps to protect themselves
           from possible identity theft. For example, consumers whose account
           information has been breached can monitor their bank or credit
           card statements for suspicious activity or close the affected
           accounts. Consumers whose personally identifiable information,
           such as SSN, has been breached can review their credit reports for
           suspicious activity or may choose to purchase a credit monitoring
           product that alerts them to changes that could indicate identity
           theft. In addition, affected consumers can place a fraud alert on
           their credit reports, which requires businesses to take certain
           identity verification steps before issuing credit.49 In some
           states, consumers can implement credit freezes, which block
           unauthorized third parties from obtaining the consumer's credit
           report or score.50 Limited information exists on the steps
           individuals actually take when notified of a breach. In the 2005
           Ponemon Institute survey of individuals that received notification
           letters, 50 percent said they did nothing, while the rest
           indicated they took actions such as monitoring their credit
           reports, canceling credit or debit cards, or closing bank
           accounts.51

           o Respecting Consumers' Right to Know. Some consumer advocates and
           others have argued that consumers have a right to know how their
           information is being handled. According to this view, basic rights
           of privacy dictate that consumers should be informed when their
           personal information has been compromised, even if the risk of
           harm is minimal. The principle that individuals should have ready
           means of learning about the use of their personal information is
           embedded in the Fair Information Practices, a set of
           internationally recognized privacy protection principles.52

           o Improving Public Awareness. Public reporting of data breaches
           may raise general awareness among consumers about the risks of
           identity theft and ways they can mitigate these risks, such as
           periodically reviewing their credit reports. In addition,
           publicity surrounding a data breach resulting from notification
           can serve to deter the use of stolen information because
           presumably the thief knows that the breach is likely being
           investigated and the stolen data are being carefully monitored.

48Such costs can be significant. For example, according to a 2006 survey,
31 companies that responded to the survey incurred an average of $98 per
record, or $2.6 million per company, in costs associated with the loss of
existing customers, recruitment of new customers, and damage to the
reputation of their brand name. Ponemon Institute, LLC, 2006 Annual Study:
Cost of a Data Breach, 2006. Due to sampling limitations, these findings
are not necessarily representative of the costs incurred by all companies
that experience breaches.

49See 15 U.S.C. S 1681c-1.

50Congressional Research Service, Identity Theft Laws: State Penalties and
Remedies and Pending Federal Bills (Washington, D.C.: Jan. 11, 2007).

51Ponemon Institute, National Survey (2005). As noted earlier, this study
may be limited by a low survey response rate.

52The Fair Information Practices were first proposed in 1973 by a U.S.
government advisory committee. A revised version was developed in 1980 by
the Organization for Economic Cooperation and Development, a group of 30
member countries that are market democracies. For more information, see
GAO, Personal Information: Agency and Reseller Adherence to Key Privacy
Principles, [40]GAO-06-421 (Washington, D.C.: Apr. 4, 2006).

Breach Notification Requirements Present a Variety of Potential Costs

According to company representatives, researchers, regulators, and others,
there are several different types of costs that may be associated with
breach notification requirements. To begin with, entities subject to
breach notification requirements may incur certain costs, regardless of
whether they actually suffer a breach, or--if they do--regardless of
whether they have to notify consumers. For example, entities may incur
costs for developing and formalizing incident response plans.

There are also the costs associated with actual notifications--potentially
including printing, postage, legal, investigative, and public relations
expenses.53 Although comprehensive data on these costs do not exist, a
2006 Ponemon Institute survey of companies experiencing a data breach
found that 31 companies that responded incurred an average of $1.4 million
per breach, or $54 per record breached, for costs related to mailing
notification letters, call center expenses, courtesy discounts or
services, and legal fees.54 Similarly, a study by Gartner Research found
that ChoicePoint spent $79 per affected account following its 2005 breach
for professional fees, legal expenses, and communications to affected
customers.55 A representative of the San Jose Medical Group told us it
spent $100,000 to send notification letters to 187,000 patients following
a data breach that occurred in 2005. Entities also may incur costs related
to staffing call centers to field inquiries from consumers about the
breach. For example, representatives of the University of California at
Berkeley told us that following a 2005 breach of 98,000 records, the
university spent $75,000 in staffing, telecommunications, and other call
center costs.

Finally, banks whose customers' account information is breached also may
incur costs for remedial steps such as canceling existing accounts or
replacing affected customers' credit or debit cards--although such steps
may not be required by the applicable breach notification requirements.
Entities experiencing a breach also often provide affected individuals
with free credit monitoring services. For example, a representative of a
large financial management company noted that offering free credit
monitoring services after a breach has become standard industry practice,
and costs, on average, between $20 and $40 per customer.

53The distinction between the costs associated with a notification
requirement versus a breach itself can be ambiguous. For example, the cost
of postage can clearly be attributed to notification, whereas legal costs
can be attributed to notification, the breach itself, or both, depending
on the circumstances.

54Ponemon Institute, Annual Study (2006). As noted earlier, due to
sampling limitations, these findings are not necessarily representative of
the costs incurred by all companies that experience breaches.

55Gartner Research, Data Protection Is Less Costly Than Data Breaches
(Stamford, Connecticut: September 16, 2005). The report, issued in 2005,
based its findings on the breach having affected 145,000 records, but
company officials later reported that 162,000 records were affected.

Challenges Exist in Complying with and Developing Breach Notification
Requirements

Officials of companies and other entities we interviewed identified
challenges such as interpreting ambiguous statutory language, identifying
and locating affected consumers, and developing effective notification
letters. In addition, policymakers face challenges in developing breach
notification requirements, particularly in setting the appropriate
standard to establish the circumstances under which consumers should be
notified.

  Complying with Notification Requirements

Companies and other entities we interviewed said they can face a number of
challenges related to complying with the breach notification requirements
in state laws or federal banking guidance. These include the following:

           o Interpreting ambiguous provisions. Entities subject to breach
           notification requirements sometimes face challenges interpreting
           certain terms or provisions of notification laws. For example, an
           information security expert told us that some laws do not
           adequately define encryption, which could refer to anything from
           simple password protection to complex coding. Similarly, federal
           banking regulators acknowledged that their institutions sometimes
           face difficulty determining whether misuse of breached information
           is "reasonably possible," such as when little information exists
           about the location of the data, the intent of a criminal who stole
           data, or the effectiveness of security features designed to render
           data inaccessible.

           o Addressing who is responsible. Notification requirements do not
           always fully address who should bear the cost of and
           responsibility for notification, particularly in cases where a
           third party is responsible for the breach. For example,
           representatives of some federal banking regulators and industry
           associations cited particular challenges associated with breaches
           of credit and debit card information by retailers. Banks that
           issue credit and debit cards compromised by a merchant that is not
           the bank's service provider are generally not required by the
           banking regulators' guidance to notify their customers, but
           nevertheless in some cases, they feel obliged to do so. Bank
           representatives with whom we spoke expressed concern that breaches
           of credit card information by third parties can adversely affect a
           bank's reputation and result in costs related to notifying
           customers and reissuing cards.

           o Identifying affected consumers. Some entities we interviewed
           said that it can be difficult to identify which consumers may have
           been affected by a breach and obtain their contact information.
           For example, one representative at a state agency involved in a
           breach told us officials were unsure what data had been downloaded
           among records that may have been accessed on 600,000 people.
           Obtaining accurate and current mailing addresses for affected
           parties also can be difficult and costly, many entities told us.
           This can be a particular problem for entities, such as merchants,
           that have breached credit card numbers but do not themselves
           possess the mailing addresses associated with those numbers.

           o Developing clear and effective notification letters. We have
           noted in the past that public notices should be useful and easy to
           understand if they are to be effective.56 However, the 2005 study
           conducted by the Ponemon Institute found that 52 percent of survey
           respondents who received a notification letter said the letter was
           not easy to understand.57 In addition, consumers might be confused
           by other mail solicitations that may resemble notification
           letters. For example, officials at one large national bank noted
           that marketing solicitations for credit monitoring services often
           are made to resemble breach notification letters, potentially
           desensitizing or confusing consumers when a true notification
           letter arrives.

           o Complying with multiple state laws. Officials of companies with
           customers in multiple states and their trade associations noted
           that they face the challenge of complying with breach notification
           requirements that vary among the states, including who must be
           notified, the level of risk that triggers a notice, the nature of
           the notification, and exceptions to the requirement. Officials of
           companies we contacted noted that it is challenging to comply with
           these multiple requirements since most breaches involve customers
           in many states.

56See [41]GAO-06-833T , Privacy: Preventing and Responding to Improper
Disclosures of Personal Information (Washington, D.C.: Jun. 8, 2006), pp.
15-18, which discusses specific elements that should be incorporated in a
breach notification.

57Ponemon Institute, National Survey (2005). As noted earlier, this study
may be limited by a low survey response rate.

  Setting an Appropriate Notification Standard

Existing state laws vary in terms of the notification standard--that is,
the event or circumstance that triggers a required notification. For
example, California has an expansive standard that requires notification
in nearly all cases where unencrypted sensitive personal data "is
reasonably believed to have been acquired by an unauthorized individual."
Other states employ a risk-based approach that incorporates into the
standard the extent to which the data are likely to be misused. The
standards vary in terms of what is required in cases where the risk of
harm is unknown. For example, Vermont requires notification unless an
entity can demonstrate that misuse of the breached data "is not reasonably
possible." In contrast, North Carolina requires notification only when it
has been determined that the breach has resulted, or is reasonably likely
to result, in illegal use of the data or creates a material risk of harm
to a consumer. As shown in figure 1, whether or not a breach is subject to
notification can depend on the specific notification standard.

Figure 1: Application of Notification Standards under Different Breach
Scenarios

Note: Figure presents hypothetical scenarios and notification standards
and is shown for illustrative purposes only.

Because of the difficulty of complying with multiple state requirements,
many companies and industry representatives have argued for a consistent
federal standard for breach notification that would preempt state
notification laws. However, the National Association of Attorneys General,
as well as some consumer and privacy groups, have expressed concern that a
federal breach notification law could weaken consumer protections if it
were to preempt stronger state laws. These groups have advocated a strong
notification standard because, they say, the link between breaches and
identity theft is not always clear and entities are not well equipped to
assess the risk of harm resulting from a given breach. As a result, too
narrow a notification standard may prevent consumers from taking action in
cases that do in fact present some risk. Also, as noted earlier, some
privacy groups and others believe that consumers have basic rights to be
notified when their personal information has been breached, no matter what
the circumstances. Moreover, they say that fears of
"overnotification"--where consumers are inundated by frequent
notifications--are unfounded, given that they are aware of no evidence of
this occurring in states that currently have strict notification
requirements.

By contrast, some representatives of the federal banking regulatory
agencies, FTC, private companies, and other experts have expressed concern
about overly expansive breach notification standards. They say that such
standards may require businesses to notify consumers about minor and
insignificant breaches. This in turn could eventually lead to
overnotification and cause consumers to spend time and money taking
proactive steps that are not necessary or, alternatively, to ignore
notices when action is warranted. In addition, businesses and federal
banking regulators have expressed concern about the financial burden that
overnotification could cause. Overly broad notification standards could
also have the effect of limiting entities' reputational incentives for
improving data security, if nearly all entities regularly issue
notifications as a result of minor breaches. Representatives of the
federal banking regulatory agencies have noted that they sought to strike
an appropriate balance with their notification standard. Their guidance
provides that, when a financial institution becomes aware of an incident
of unauthorized access to sensitive customer information, the institution
should conduct a reasonable investigation to promptly determine the
likelihood that the information has been or will be misused.58 If the
institution determines that misuse of the information has occurred or is
reasonably possible, it should notify affected customers as soon as
possible. The guidance is intended to provide notice to customers only
when there is a reasonable expectation of misuse.59

5812 C.F.R. Pt. 30, App. B, Supp. A S III(A); 12 C.F.R. Pt. 208, App. D-2,
Supp. A S III(A); 12 C.F.R. Pt. 225, App. F, Supp. A S III(A); 12 C.F.R.
Pt. 364, App. B, Supp. A S III(A); 12 C.F.R. Pt. 570, App. B, Supp. A S
III(A); and 12 C.F.R. Pt. 748, App. B S III(A).

Similarly, the guidance for federal agencies developed by the President's
Identity Theft Task Force recommended that if an agency experiences a
breach, it should analyze the risk of identity theft and tailor its
response--which may include notifying individuals--to the nature and scope
of the risk presented. The guidance noted that such a risk assessment can
minimize the potentially significant costs of notification where little
risk exists. The task force's April 2007 strategic plan recommended the
development of a national standard requiring all entities that maintain
sensitive consumer information, in both the public and private sectors, to
provide notice to consumers and law enforcement in the event of a breach.
As with its guidance to federal agencies, the task force recommended that
the standard be risk based to provide notice when consumers face a
significant risk of identity theft but to avoid excessive notification.

As we have noted in the past, care is needed in defining appropriate
criteria for data breaches that merit notification.60 The frequency of
data breaches identified in this report suggests that a national breach
notification requirement may be beneficial, in large part because of its
role in further encouraging entities to improve their data security
practices. However, because breaches vary in the risk they present, and
because most breaches have not resulted in detected incidents of identity
theft, a notification that is risk based appears appropriate. Should
Congress choose to enact a federal breach notification requirement, use of
the risk-based approaches that the federal banking regulators and the
President's Identity Theft Task Force advocate could avoid undue burden on
organizations and unnecessary and counterproductive notifications to
consumers.

59The guidance states that institutions should notify their primary
federal regulator as soon as possible when the institution becomes aware
of an incident involving unauthorized access to or use of sensitive
customer information, even for incidents that may not warrant customer
notification. Banking regulators told us they review institutions'
response programs as part of their supervisory procedures and, in many
cases, work with institutions as they respond to specific incidents to
ensure their actions are in accordance with the guidance. See 12 C.F.R.
Pt. 30, App. B, Supp. A S II(A)(1)(b); 12 C.F.R. Pt. 208, App. D-2, Supp.
A S II(A)(1)(b); 12 C.F.R. Pt. 225, App. F, Supp. A S II(A)(1)(b); 12
C.F.R. Pt. 364, App. B, Supp. A S II(A)(1)(b); 12 C.F.R. Pt. 570, App. B,
Supp. A S II(A)(1)(b); and 12 C.F.R. Pt. 748, App. B S II(A)(1)(b).

60GAO, Personal Information: Key Federal Privacy Laws Do Not Require
Information Resellers to Safeguard All Sensitive Data, [42]GAO-06-674
(Washington, D.C.: Jun. 26, 2006) and [43]GAO-06-833T .

Agency Comments

We provided a draft of this report to FTC, which provided technical
comments that were incorporated in this report as appropriate. In
addition, we provided selected portions of the draft to the Board of
Governors of the Federal Reserve System, the Department of Justice, DHS,
FDIC, the National Credit Union Administration, the Office of the
Comptroller of the Currency, the Office of Thrift Supervision, the Social
Security Administration, and USPIS, and also incorporated their technical
comments as appropriate.

As agreed with your offices, unless you publicly announce the contents of
this report earlier, we plan no further distribution of it until 30 days
from the date of this letter. We will then send copies of this report to
the Chairman, House Committee on Financial Services; the Chairman and
Ranking Member, Senate Committee on Banking, Housing, and Urban Affairs;
the Chairman and Ranking Member, Senate Committee on the Judiciary; the
Chairman and Ranking Member, House Committee on the Judiciary; the
Chairman and Vice Chairman, Senate Committee on Commerce, Science, and
Transportation; and the Chairman and Ranking Member, House Committee on
Energy and Commerce. We will also send copies to the Chairman of the Board
of Governors of the Federal Reserve System, the Attorney General, the
Secretary of the Department of Homeland Security, the Chairman of the
Federal Deposit Insurance Corporation, the Chairman of the Federal Trade
Commission, the Chairman of the National Credit Union Administration, the
Comptroller of the Currency, the Director of the Office of Thrift
Supervision, the Commissioner of the Social Security Administration, and
the Postmaster General and Chief Executive Officer of the U.S. Postal
Service. We will also make copies available to others upon request. In
addition, the report will be available at no charge on GAO's Web site at
[44]http://www.gao.gov .

If you or your staff have any questions about this report, please contact
me at (202) 512-8678 or [email protected]. Contact points for our Offices of
Congressional Relations and Public Affairs may be found on the last page
of this report. GAO staff who made key contributions to this report are
listed in appendix II.

David G. Wood
Director, Financial Markets and Community Investment

Appendix I: Scope and Methodology

Our report objectives were to examine (1) what is known about the
incidence and circumstances of breaches of sensitive personal information;
(2) what information exists on the extent to which breaches of sensitive
personal information have resulted in identity theft; and (3) the
potential benefits, costs, and challenges associated with breach
notification requirements. We use the term "data breach" to refer to the
unauthorized or unintentional exposure, disclosure, or loss of sensitive
personal information by a company, government agency, university, or other
public or private entity. Our scope was limited to breaches involving
personal data, including financial data, that could be used to commit
identity theft or other related harm, and we excluded breaches involving
other types of sensitive data, such as medical records or proprietary
business information. For the purposes of this report, the term "identity
theft" is used broadly to refer to both fraud on existing accounts and the
unauthorized creation of new accounts.

To address all three objectives, we conducted a literature search of
relevant articles, reports, and studies. We also collected and analyzed
documents from, and interviewed, officials of government agencies that
investigate and track data breaches, including the Federal Trade
Commission, the Department of Homeland Security, the Department of
Justice, the U.S. Postal Inspection Service, and the Social Security
Administration. We also interviewed staff at the five federal banking
regulators--the Board of Governors of the Federal Reserve System, the
Federal Deposit Insurance Corporation, the Office of the Comptroller of
the Currency, the Office of Thrift Supervision, and the National Credit
Union Administration. In addition, we spoke with representatives of the
National Association of Attorneys General and organizations that address
consumer protection and privacy issues, including Consumers Union,
Electronic Privacy Information Center, Privacy Rights Clearinghouse,
Attrition, and the Identity Theft Resource Center. We also spoke with
three academic researchers who study issues related to data breaches and
notification and an attorney who helps companies address data privacy and
security issues. In addition, we reviewed studies on data breaches
conducted by private and nonprofit research organizations, including the
Ponemon Institute, ID Analytics, and Javelin Strategy and Research. We
interviewed the studies' authors and took other steps to ensure that the
data and methodologies were sufficiently reliable for our purposes. We
also spoke with representatives of the California Office of Privacy
Protection and its advisory group and reviewed the office's recommended
practices for notification.

To address the first objective on the incidence and circumstances of data
breaches, we reviewed lists of news media-reported data breaches that are
compiled and maintained by three private research and advocacy
organizations--Privacy Rights Clearinghouse, Attrition, and the Identity
Theft Resource Center. We analyzed the three independent lists to create a
single, nonduplicative list of data breaches that had been reported in the
news media from January 2005 through December 2006. We took measures to
ensure the lists were of sufficient quality for our purposes, including
spot checking selected data and interviewing representatives of the three
organizations on their methodologies. The Privacy Rights Clearinghouse,
Attrition, and Identity Theft Resource Center lists contained 436, 453,
and 462 breaches, respectively, for the time period we analyzed. Of the
572 breaches they collectively compiled, 59 percent appeared on all three
lists, 19 percent appeared on two, and 22 percent appeared on one. Our
analysis was based on the lists as they stood on February 15, 2007; these
data may have changed because the lists are occasionally updated when the
compilers learn of new breaches that may have occurred in the past.

We also collected available data from federal law enforcement agencies on
the breaches they have investigated in recent years. In addition, the five
federal banking regulators provided, at our request, data on the breaches
of which they have been notified by the institutions they supervise. These
data varied in usefulness and comprehensiveness because of the regulators'
differing methods of counting and tracking breaches and maintaining data
on them. We also gathered data from two states, New York and North
Carolina, which were selected because they were two large states that
maintain centralized information on breaches. Further, we obtained
available data from industry and trade associations representing key
sectors--such as financial services, retail sales, higher education,
hospitals, and information services--that have experienced data breaches.
We also collected information on breaches experienced by federal agencies
compiled by the House Government Reform Committee and the U.S. Computer
Emergency Readiness Team, a component of the Department of Homeland
Security.

To address the second objective, we selected for more detailed examination
the 24 largest (in terms of number of records breached) data breaches
reported in the news media from January 2000 through June 2005. We
selected these breaches in August 2006 using the lists maintained by
Privacy Rights Clearinghouse and Identity Theft Resource Center, as well
as a similar compilation of breaches collected by the Congressional
Research Service. We were not aware of the Attrition list at the time we
made our selection. For each of these breaches, we reviewed news reports
as well as publicly available documents such as testimonies and criminal
indictments. We also conducted interviews, where possible, with
representatives of the entities that experienced the breach and law
enforcement agencies that investigated the breach. We identified those
cases where this information collectively indicated that the breach
appeared to have resulted in some form of identity theft. Ultimately, the
determination of whether particular conduct violated a law prohibiting
identity theft would be a matter of law for the courts. We did not
directly contact individuals whose data had been affected by the breaches
because of privacy concerns and because we did not have a systematic means
of identifying them. We also reviewed five breaches that reportedly
involved federal agencies--the Navy; the Internal Revenue Service; the
Federal Deposit Insurance Corporation; the National Park Service; and the
Department of Justice. These were selected to represent breaches that
included different causes, types of data, and involvement by third-party
vendors.

To examine the potential benefits, costs, and challenges associated with
breach notification requirements, we reviewed the federal banking
regulators' proposed and final guidance related to breach notification,
and interviewed representatives of each agency regarding their
consideration of potential costs, benefits, and challenges during
development of the guidance. Further, we reviewed the strategic plan and
other documents issued by the President's Identity Theft Task Force. In
addition, we conducted a review of the effects of California's breach
notification law. We interviewed representatives of, and gathered
information from, seven organizations to learn about their experiences
complying with California's breach notification law. These organizations
were selected to represent a range of organization sizes and industry
sectors. We also interviewed representatives of the California State
Information Security Office, California State Assembly, California Office
of Privacy Protection, and California Bankers Association.

We conducted our review from August 2006 through April 2007 in accordance
with generally accepted government auditing standards.

Appendix II: GAO Contact and Staff Acknowledgments

GAO Contact

David G. Wood, (202) 512-8678 or [email protected]

Staff Acknowledgments

In addition to the contact named above, Jason Bromberg, Assistant
Director; Randy Fasnacht; Marc Molino; Kathryn O'Dea; Carl Ramirez; Linda
Rego; Barbara Roesmann; and Winnie Tsen made key contributions to this
report.

(250295)

GAO's Mission

The Government Accountability Office, the audit, evaluation and
investigative arm of Congress, exists to support Congress in meeting its
constitutional responsibilities and to help improve the performance and
accountability of the federal government for the American people. GAO
examines the use of public funds; evaluates federal programs and policies;
and provides analyses, recommendations, and other assistance to help
Congress make informed oversight, policy, and funding decisions. GAO's
commitment to good government is reflected in its core values of
accountability, integrity, and reliability.

Obtaining Copies of GAO Reports and Testimony

The fastest and easiest way to obtain copies of GAO documents at no cost
is through GAO's Web site ( [45]www.gao.gov ). Each weekday, GAO posts
newly released reports, testimony, and correspondence on its Web site. To
have GAO e-mail you a list of newly posted products every afternoon, go to
[46]www.gao.gov and select "Subscribe to Updates."

Order by Mail or Phone

The first copy of each printed report is free. Additional copies are $2
each. A check or money order should be made out to the Superintendent of
Documents. GAO also accepts VISA and Mastercard. Orders for 100 or more
copies mailed to a single address are discounted 25 percent. Orders should
be sent to:

U.S. Government Accountability Office 441 G Street NW, Room LM Washington,
D.C. 20548

To order by Phone: Voice: (202) 512-6000
TDD: (202) 512-2537
Fax: (202) 512-6061

To Report Fraud, Waste, and Abuse in Federal Programs

Contact:

Web site: [47]www.gao.gov/fraudnet/fraudnet.htm
E-mail: [48][email protected]
Automated answering system: (800) 424-5454 or (202)
512-7470

Congressional Relations

Gloria Jarmon, Managing Director, [49][email protected] (202) 512-4400 U.S.
Government Accountability Office, 441 G Street NW, Room 7125 Washington,
D.C. 20548

Public Affairs

Paul Anderson, Managing Director, [50][email protected] (202) 512-4800
U.S. Government Accountability Office, 441 G Street NW, Room 7149
Washington, D.C. 20548

[51]www.gao.gov/cgi-bin/getrpt?GAO-07-737 .

To view the full product, including the scope
and methodology, click on the link above.

For more information, contact David G. Wood at (202) 512-8678 or
[email protected].

Highlights of [52]GAO-07-737 , a report to congressional requesters

June 2007

PERSONAL INFORMATION

Data Breaches Are Frequent, but Evidence of Resulting Identity Theft Is
Limited; However, the Full Extent Is Unknown

In recent years, many entities in the private, public, and government
sectors have reported the loss or theft of sensitive personal information.
These breaches have raised concerns in part because they can result in
identity theft--either account fraud (such as misuse of credit card
numbers) or unauthorized creation of new accounts (such as opening a
credit card in someone else's name). Many states have enacted laws
requiring entities that experience breaches to notify affected
individuals, and Congress is considering legislation that would establish
a national breach notification requirement.

GAO was asked to examine (1) the incidence and circumstances of breaches
of sensitive personal information; (2) the extent to which such breaches
have resulted in identity theft; and (3) the potential benefits, costs,
and challenges associated with breach notification requirements. To
address these objectives, GAO reviewed available reports on data breaches,
analyzed 24 large data breaches, and gathered information from federal and
state government agencies, researchers, consumer advocates, and others.

[53]What GAO Recommends

This report contains no recommendations.

While comprehensive data do not exist, available evidence suggests that
breaches of sensitive personal information have occurred frequently and
under widely varying circumstances. For example, more than 570 data
breaches were reported in the news media from January 2005 through
December 2006, according to lists maintained by private groups that track
reports of breaches. These incidents varied significantly in size and
occurred across a wide range of entities, including federal, state, and
local government agencies; retailers; financial institutions; colleges and
universities; and medical facilities.

The extent to which data breaches have resulted in identity theft is not
well known, largely because of the difficulty of determining the source of
the data used to commit identity theft. However, available data and
interviews with researchers, law enforcement officials, and industry
representatives indicated that most breaches have not resulted in detected
incidents of identity theft, particularly the unauthorized creation of new
accounts. For example, in reviewing the 24 largest breaches reported in
the media from January 2000 through June 2005, GAO found that 3 included
evidence of resulting fraud on existing accounts and 1 included evidence
of unauthorized creation of new accounts. For 18 of the breaches, no clear
evidence had been uncovered linking them to identity theft; and for the
remaining 2, there was not sufficient information to make a determination.

Requiring affected consumers to be notified of a data breach may encourage
better security practices and help mitigate potential harm, but it also
presents certain costs and challenges. Notification requirements can
create incentives for entities to improve data security practices to
minimize legal liability or avoid public relations risks that may result
from a publicized breach. Also, consumers alerted to a breach can take
measures to prevent or mitigate identity theft, such as monitoring their
credit card statements and credit reports. At the same time, breach
notification requirements have associated costs, such as expenses to
develop incident response plans and identify and notify affected
individuals. Further, an expansive requirement could result in
notification of breaches that present little or no risk, perhaps leading
consumers to disregard notices altogether. Federal banking regulators and
the President's Identity Theft Task Force have advocated a notification
standard--the conditions requiring notification--that is risk based,
allowing individuals to take appropriate measures where the risk of harm
exists, while ensuring they are only notified in cases where the level of
risk warrants such action. Should Congress choose to enact a federal
notification requirement, use of such a risk-based standard could avoid
undue burden on organizations and unnecessary and counterproductive
notifications of breaches that present little risk.

References

Visible links
  36. http://www.gao.gov/cgi-bin/getrpt?GAO-07-657
  37. http://www.gao.gov/cgi-bin/getrpt?GAO-05-710
  38. http://www.gao.gov/cgi-bin/getrpt?GAO-02-363
  39. http://www.gao.gov/cgi-bin/getrpt?GAO-06-674
  40. http://www.gao.gov/cgi-bin/getrpt?GAO-06-421
  41. http://www.gao.gov/cgi-bin/getrpt?GAO-06-833T
  42. http://www.gao.gov/cgi-bin/getrpt?GAO-06-674
  43. http://www.gao.gov/cgi-bin/getrpt?GAO-06-833T
  44. http://www.gao.gov/
  45. http://www.gao.gov/
  46. http://www.gao.gov/
  47. http://www.gao.gov/fraudnet/fraudnet.htm
  48. mailto:[email protected]
  49. mailto:[email protected]
  50. mailto:[email protected]
  51. http://www.gao.gov/cgi-bin/getrpt?GAO-07-737
  52. http://www.gao.gov/cgi-bin/getrpt?GAO-07-737
*** End of document. ***