The FBI calculated the price tag by extrapolating results from a survey of 2,066 organizations. The survey, released Thursday, found that 1,324 respondents, or 64 percent, suffered a financial loss from computer security incidents over a 12-month period.
The average cost per company was more than $24,000, with the total cost reaching $32 million for those surveyed.
Often survey results can be skewed, because poll respondents are more likely to answer when they have experienced a problem. So, when extrapolating the survey results to estimate the national cost, the FBI reduced the estimated number of affected organizations from 64 percent to a more conservative 20 percent.
"This would be 2.8 million U.S. organizations experiencing at least one computer security incident," according to the . "With each of these 2.8 million organizations incurring a $24,000 average loss, this would total $67.2 billion per year."
By comparison, telecommunication fraud losses are about only $1 billion a year, according to the U.S. Secret Service. Also, the overall cost to Americans of, according to Javelin Strategy & Research.
Other surveys have attempted to put a dollar amount on cybersecurity damages in the past, but the FBI believes its estimate is the most accurate because of the large number of respondents, said Bruce Verduyn, the special agent who managed the survey project.
"The data set is three or four times larger than in past surveys," he said. "It is obviously a staggering number, but that is the reality of what we see."
Responding to was most costly, followed by , financial fraud and network intrusion, according to the survey. Respondents spent nearly $12 million to deal with virus-type incidents, $3.2 million on theft, $2.8 million on financial fraud and $2.7 million on network intrusions.
These figures do not include much of the staff, technology, time and software employed to prevent security incidents, Verduyn said. Also, losses to individuals who are victims of computer crime or victims in other countries are not included, he said.
The FBI's next fiscal year, for which budgets must be reviewed and approved, begins Oct. 1. Protecting the U.S. against high technology crimes is third on the agency's list of priorities.
Defenses in place
Survey respondents use a variety of security products for protection. Antivirus software is almost universally used, with 98.2 percent of respondents stating they use it. Firewalls follow in second place, with 90.7 percent, and anti-spyware and antispam are each used by about three-quarters of respondents, according to the survey.
The results mean that close to one in 10 organizations does not have a hardware or software firewall. Or perhaps they don't know they have one--the Windows Firewall in Windows XP, for example. "Some are very small businesses that should have that technology, but they don't," Verduyn explained.
and --both relatively new security technologies--were used only by 4 percent and 7 percent of survey respondents, respectively. Intrusion prevention or detection systems were used by 23 percent and VPNs, or virtual private networks, by 46 percent.
Organizations were attacked despite use of security products, with nine out of 10 respondents saying they experienced a security incident. In fact, the most common attacks aligned with the most commonly used defenses. Computer viruses, worms or Trojan horses plagued 84 percent of respondents, 80 percent reported spyware trouble, and 32.9 percent said attackers were probing their systems using.
Not all threats came from outside the organization. More than 44 percent of the survey respondents reported intrusions from within the company. "Companies may be unaware of the internal potential for computer security incidents," Verduyn said. He recommends applying policies and procedures to thwart attacks from the inside.
The FBI surveyed companies in Iowa, Nebraska, New York and Texas. Companies older than three years, with more than five employees and with more than $1 million in revenue were asked to participate. Survey participants were asked to provide their responses by the end of July 2005, with their answers covering the previous 12-month period.