The Javelin Identity Fraud Survey Report provides consumers and businesses an in-depth and comprehensive examination of identity fraud in the United States. Its purpose is to help readers understand the causes and incidence rates of identity fraud and the success rates of methods used for its prevention, detection and resolution.
This report builds on Javelin’s annually published Identity Fraud Survey Report and the Federal Trade Commission’s 2003 Identity Theft Survey Report.
Survey Questionnaire
The set of questions and underlying methodology used for this report were identical to or highly similar to the 2008, 2007, 2006, 2005, 2004 and 2003 surveys. This allows the ability to provide longitudinal trends on various subjects, such as incidence rates and detection methods.
In addition, to more deeply explore the significance of past responses and to identify the lasting impact of identity fraud on victims, Javelin added a discrete number of new questions. These expanded on the behaviors of consumers after personal information was compromised and/or misused.
Some questions from the previous surveys were modified to improve the accuracy of estimates. Questions used to measure the fraud amounts, consumer cost and resolution hours previously provided ranges of responses. In the 2008, 2007 and 2006 reports, these questions were changed to collect exact amounts.
Survey Respondents
In all, 5,000 consumers, representative of the U.S. population, were interviewed via a standardized 50-question telephone survey to develop accurate and actionable insight into this pervasive and costly crime.
The polling yielded interviews with 828 fraud victims. After weighting the responses to standardize them to national demographics*, the 2009 survey’s computed number of victims interviewed was 703. For comparison, the numbers of victims in previous years are provided below, before and after weighting:
Survey Data Collection
Javelin employed Harris Interactive Service Bureau (HISB) for this survey’s data collection. HISB, one of the nation’s leading data collection providers, is recognized as a reputable data collection service firm with almost 50 years of experience in the industry. Data for this survey were collected by Harris Interactive Service Bureau on behalf of Javelin. HISB was responsible for collecting the data and Javelin was responsible for the survey design, data weighting, data analysis and reporting. Previous studies employed Discovery and Synovate for all phases of data collection using computer-assisted telephone interviewing (CATI) via random-digit Dialing (RDD).
The study was conducted using interviews administered by telephone with 5,000 U.S. adults over age 18 and a sample that is representative of the U.S. census demographics distribution. Data collection began Sept. 19, 2009 and ended Nov. 29, 2009.
Margin of Error
For questions answered by all 5,000 respondents, the maximum margin of sampling error is +/- 1.4% at the 95% confidence level. For questions answered by all 703 identity fraud victims, the maximum margin of sampling error is +/- 3.7% at the 95% confidence level. For questions answered by a proportion of all identity fraud victims, the maximum margin of sampling error varies and is greater than +/- 3.7% at the 95% confidence level.
Categorizing Fraud
With one exception, this report continues to classify fraud within the three categories originally defined by the FTC. For 2005 and beyond, debit card fraud has been re-categorized as existing card accounts fraud** instead of existing non-card accounts fraud.*** Javelin believes that this change reflects a more accurate representation of debit card fraud because much of its means of compromise, fraudulent use and detection methods parallel those of credit cards.
The categories of fraud are listed below from least to most serious:
- Existing card accounts: This category includes both the account numbers and/or the actual cards for existing credit and card-linked debit accounts. Prepaid cards were added for 2007 and subsequently removed due to extremely low incidence.
- Existing non-card accounts: This category includes existing checking and savings accounts, and existing loans, insurance, telephone and utilities accounts.
- New accounts and other frauds: This category includes new accounts or loans for committing theft, fraud or other crimes using the victim's personal information.
Many victims experience identity fraud within more than one of these categories. In reporting the overall incidence rates of the three categories or types of accounts, the victims of crimes to more than one type of account are categorized based on the most serious (as designated by the FTC) problem reported. Thus, victims who reported that new accounts had been opened using their information and also that their existing credit cards had been misused would be placed in the new accounts and other frauds classification, not in the existing card accounts classification. This categorization is applicable only for reporting the rates of the three types of fraud.
Reporting Years of Findings
This report labels longitudinal findings according to the year the data was collected, in contrast to previous years when longitudinal findings were labeled according to the year the report was published. Most notably, the current year of the data in this report is labeled as 2009 as the data collection was finished in November 2009.
Calculations
Comparing Annual Numbers
To create a more accurate understanding of the costs of identity fraud, Javelin has departed from using ranged brackets (e.g., $0 to $50) to measure the fraud amount, consumer cost and hours spent resolving fraud. This change in methodology leads to a lower estimate of the average fraud amount.
By converting the data collected into the 2005 ‘‘bracketed’’ year’s methodology, Javelin has found that the means derived from the former bracketed methodology overstate the average losses by more than 20% (with the exception of the final category, where the bracketed fraud amount decreased the average fraud amount of the victims by 55%). By adopting this change, Javelin has created a more accurate picture of the impact of identity fraud in the U.S.
To allow for comparisons to previous years, it is necessary to re-bracket the actual amounts and this is what has been done. In the future, access to these actual (versus bracketed) numbers will allow for comparisons of like figures using the more accurate actual numbers.
Due to rounding errors, the percentages on graphs add up to 100% plus or minus one percent.
To assure consistency in comparing year-to-year changes, historical figures for average fraud amounts were adjusted for inflation by using the Consumer Price Index (CPI). 2008 survey numbers were adjusted by 0.4%, 2007 survey numbers were adjusted by 4.3%, 2006 survey numbers were adjusted by 7.2%, 2005 survey numbers were adjusted by 10.7%, 2004 survey numbers were adjusted by 14.4% and 2003 numbers were adjusted by 17.5% to normalize the value of currency to 2009 dollars. ****
Data Smoothing Techniques
2005, 2006, 2007, 2008, and 2009 total dollar cost estimates have been smoothed by taking a moving three-year average of the victim’s fraud amount. Time series data smoothing techniques are used to eliminate “noise” created by random data fluctuations and uncover real trends. Fraud data by state is estimated using a three-year moving average.
Calculating Mean and Median Values
Where responses pertained to a range in value, e.g., “one day to less than one week,” the midpoint of the range, rounded up to the nearest whole unit , was used to calculate the median or mean value.
Example: If the response selected for fraud amount was one day to less than one week, the assigned value would be the median of one day and seven days, inclusive, or 4 days.
Deviation from FTC and 2003 Methodology and Reporting
When reporting victims’ average financial damages or resolution times in dollars or hours, the entire amount of damages or losses are placed into every type of fraud the victims suffered. For example, for a victim who reports that a total of $100 is obtained for both new accounts and other frauds category and existing card accounts, the $100 is counted in both categories. This method of reporting costs by types of fraud will not change the overall total costs of fraud across all three categories, but the average amount of dollars or time associated within the three types of fraud should not be summed because there will be overlapping amounts.
In the 2003 report, responses to the new accounts and other frauds question (Q13) were modified based on respondents’ subsequent answers to question 18. 2009’s question 21 is slightly modified from 2003’s, thus avoiding the possibility of needing to adjust responses to question 13 to maintain the longitudinal integrity. 2009’s responses to question 21 are reported as they were reported by the victims.
Secondly, 2009’s detection time question (Q22) is categorized differently from the 2003 study. While the 2003 study provided 13 answers from which victims could choose, the 2006 study contained only nine such responses. Javelin merged similar response categories that contained few replies in 2003 into single categories, allowing the data to be cross-tabbed with larger numbers and fewer categories for a more robust calculation.
On several other questions, longitudinal comparisons are performed with numbers that Javelin calculated using 2003 raw data instead of 2003 reported figures. This was done to avoid inserting rounding errors or methodology differences.
Contributing Organizations
The survey was in part made possible by Fiserv, Intersections, Inc., Wells Fargo, and Identity Theft Assistance Center (ITAC). To preserve the project’s independence and objectivity, the sponsors of this project were not involved in the tabulation, analysis or reporting of final results.
* Responses were adjusted to be nationally representative based on respondents’ age, gender, income level, and race/ethnicity as reported by U.S. Census on Dec. 11, 2006. http://www.census.gov/prod/www/abs/statab2001_2005.html accessed Dec. 11, 2009.
** Formerly titled “Existing Credit Card Accounts” in the 2003 and 2005 reports
*** Formerly titled “Existing Non-Credit Card Accounts” in the 2003 and 2005 reports
**** Dollar costs were adjusted using the Consumer Price Index (CPI-U, Base 1982-1984=100) issued by the Bureau of Labor Statistics, ftp://ftp.bls.gov/pub/special.requests/cpi/cpiai.txt accessed Dec. 3, 2009.