The Story of the Social Security Number

Introduction

The Social Security number (SSN) was created in 1936 for the sole purpose of tracking the earnings histories of U.S. workers, for use in determining Social Security benefit entitlement and computing benefit levels. Since then, use of the SSN has expanded substantially. Today the SSN may be the most commonly used numbering system in the United States. As of December 2008, the Social Security Administration (SSA) had issued over 450 million original SSNs, and nearly every legal resident of the United States had one. The SSN's very universality has led to its adoption throughout government and the private sector as a chief means of identifying and gathering information about an individual.

How did the SSN come to be, and why has it become an unofficial national identifier? This article explores the history and meaning of the SSN and the Social Security card, along with SSA's SSN master data file, generally known as the Numident. The article also traces how use of the SSN has expanded since its introduction and the steps SSA has taken to enhance the integrity of the SSN process.

Crafting the SSN

At its inception, the SSN's only purpose was to uniquely identify U.S. workers, enabling employers to submit accurate reports of covered earnings for use in administering benefits under the new Social Security program. That is still the primary purpose for the SSN.

However, creating the SSN scheme and assigning SSNs to U.S. workers was no easy task. Passage of the Social Security Act in August 1935 set in motion a huge effort to build the infrastructure needed to support a program affecting tens of millions of individuals. Many said the task was impossible (SSA 1952; SSA 1965, 26). Employers were to begin to deduct payroll taxes from worker's wages in January 1937, giving the agency little time to establish the SSN process.¹ Besides clarifying program policy, the agency needed to hire and train employees (7,500 by March 1938), set up facilities, develop public education programs, and create an earnings-tracking system (Corson 1938, 6).

Establishing the Social Security infrastructure was impeded for 3½ months by the lack of funds due to a filibuster of the 1936 Deficiency Bill (a government-wide appropriation bill similar to current Omnibus Budget Reconciliation bills) by Senator Huey Long (D–LA). The Roosevelt administration circumvented this obstacle by engineering a Works Progress Administration (WPA) allotment of $112,610 from the Department of Labor and by borrowing staff from the demobilizing National Recovery Administration, the Federal Emergency Relief Administration, and the National Youth Administration. On February 9, 1936, Congress finally passed the deficiency bill containing the fiscal year 1936 appropriation for the Social Security Board (precursor of the SSA), and Roosevelt signed it on February 11. As late as March 15, 1936, there were still only five employees of the Social Security Board's Bureau of Old-Age Benefits—including the director and his assistant (McKinley and Frase 1970, 18, 28, and 49).

Additionally, the U.S. Supreme Court declared the Agricultural Adjustment Act invalid in January 1936, raising the possibility that the Social Security Act might also be declared unconstitutional (McKinley and Frase 1970, 22–23; SSA 1952). It would not be until May 1937 that a series of Supreme Court decisions cemented the constitutionality of the Social Security Act (SSA n.d. a).

One of the first steps in administering the Social Security Act was to devise a means to track the earnings of each individual, as Social Security benefit computations consider a worker's earnings from 1937 on.

Why didn't the Social Security Board just use an individual's name and address as the identifier? The deficiency of such a scheme was already well known. A 1937 publication recounts, "A recent news account states that the Fred Smiths of New York City have had so much trouble in being identified by their creditors, the courts, and even their friends, that they have joined together in forming the 'Fred Smiths, Incorporated,' to serve as a clearing house for their identification problems." Some government agencies, such as the U.S. War and Navy Departments, the Veterans Administration (for paying pensions and for adjusted compensation certificates), and the Post Office Department (for Postal Savings depositors) used fingerprints for identification. However, the use of fingerprints was associated in the public mind with criminal activity, making this approach undesirable (Wyatt and Wandel 1937, 45–47). A numbering scheme was seen as the practical alternative. Thus, the employer identification number (EIN) and the SSN were created.

Today we take the 9-digit composition of the SSN as a given, but in 1935 and 1936 many other schemes were considered. In early November 1935, the Social Security Board adopted an identifier composed of 3 alphabetic characters representing geographic areas and 5 numeric characters. However, the Board made this decision without consulting other federal agencies. The U.S. Employment Service (USES), the Census Bureau, the Central Statistical Board, and the Bureau of Labor Statistics all used numeric symbols without alphabetic characters since most standard statistical machines used this scheme. With alphabetic symbols, these agencies, as well as many private companies, would have had to buy new machines. Only two companies manufactured tabulating machines using a combination of alphabetic and numeric characters, and the government had previously brought suit against them under the Sherman Antitrust Act for dividing market territory between them (McKinley and Frase 1970, 320–322).

The Board called a meeting of all interested agencies to discuss the numeration issues.² In a November 1935 report, a subcommittee of this interdepartmental group proposed three alternatives:

• a 9-digit number consisting of a 4-digit serial number, a 2-digit year of birth indicator, and a 3-digit number indicating the geographic area of registration;
• an 8-digit number with a 5-digit serial number and a 3-digit geographic indicator; or
• a 7-character version consisting of 4 digits and 3 alphabetic characters (McKinley and Frase 1970, 322).

On December 17, 1935, the Board approved the 9-digit option (McKinley and Frase 1970, 323). The Board planned to use the year one attained age 65 as part of the SSN, thinking that once an individual attained age 65, the SSN would be reassigned to someone else. But at a meeting on January 23, 1936, the unemployment compensation delegates objected to the use of digits to signify age because they thought a number of workers would falsify their age. As a result, a new scheme adopted by the Board on February 14 consisted of a 3-digit area code, a 2-digit month of birth, and a 4-digit serial number.

Finally, on June 2, 1936, the Board decided to keep the 9-digit scheme, although using the fourth and fifth digits to represent the month of birth was abandoned. Instead, those two digits would be a "group number" that could be used to maximize the utility of mechanical equipment and to verify the accuracy of punch cards. This scheme would permit the prenumbering of registration forms and was capable of expansion to nearly 1 billion accounts (McKinley and Frase 1970, 342–344). The numbering scheme would also facilitate storing the applications since the agency's files were organized by region as well as alphabetically.

Deconstructing the SSN

As a result of the June 1936 decision, the current SSN is composed of three parts:

• The first three digits are the area number
• The next two digits are the group number
• The final four digits are the serial number

Area Number

The 3-digit area number is assigned by geographic region. In 1936 the Social Security Board planned eventually to use area numbers to redistribute work to its 12 regional centers to serve workers in those areas. One or more area numbers were allocated to each state based on the anticipated number of SSN issuances in the state.³ Prior to 1972, the numbers were issued to local offices for assignment to individuals; it was thought this would capture information about the worker's residence. So, until 1972, the area number represented the state in which the card was issued. (Barron and Bamberger 1982, 29).

Generally, area numbers were assigned in ascending order beginning in the northeast and then moving westward. For the most part, people on the east coast have the lowest area numbers and those on the west coast have the highest area numbers. However, area numbers did not always reflect the worker's residence. During the initial registration in 1936 and 1937, businesses with branches throughout the country had employees return their SS-5 Application for Account Number to their national headquarters, so these SSNs carried the area number where the headquarters were located. As a result, the area numbers assigned to big cities, such as New York, Philadelphia, Boston, and Chicago, were used for workers in many other parts of the country (McKinley and Frase 1970, 373). Also, a worker could apply in person for a card in any Social Security office, and the area number would reflect that office's location, regardless of the worker's residence.

Since 1972, when SSA began assigning SSNs and issuing cards centrally from Baltimore, MD, the area number has been assigned based on the ZIP code of the mailing address provided on the application for the original Social Security card. The applicant's mailing address may not be the same as the place of residence.

Some exceptions to the general east-to-west, ascending-order area numbering scheme exist:

• Sequence 700 through 728 was assigned to railroad workers until July 1963.
• 586 was divided among American Samoa, Guam, the Philippines, Americans employed abroad by American employers and, from 1975 to 1979, Indochinese refugees.
• 580 was assigned to Puerto Rico and the U.S. Virgin Islands; sequences 581 through 584 and 596 through 599 were also assigned to Puerto Rico.
• Sequence 577 through 579 was assigned to the District of Columbia.
• Sequences 587 through 588 and 589 through 595 were assigned to Mississippi and Florida, respectively, for use after those states exhausted their initial area number allotments.
• Sequence 729 through 733 has been allocated to the Department of Homeland Security (DHS) for SSNs issued through the Enumeration at Entry (EaE) program, described below.
• No SSNs with an area number in the 800s or 900s, or with a 000 area number, have been assigned.
• No SSNs with an area number of 666 have been or will be assigned.

SSA has many years' worth of potential SSNs available for future assignment. However, because of population shifts, SSA now faces an imbalance in the geographic allocation of area numbers. Some states have a current allocation of SSNs that will last for many years, while others have a pending shortage. As a result, given present rates of assignment and existing geographic allocations, several states currently have fewer than 10 years' worth of SSNs available for assignment.

In a July 3, 2007, Federal Register notice, SSA solicited public comment on a proposal to change the way SSNs are assigned (SSA 2007b). Under this proposal, SSA would randomly assign SSNs from the remaining pool of available numbers, and the first three digits would no longer have any geographic significance. SSA contends that doing so would ensure a reliable supply of SSNs for years to come, and would also reduce opportunities for identity theft and SSN fraud and misuse. SSA plans additional discussion with other government entities and the private sector before implementing any change.

Group Number

The group number (the fourth and fifth digits of the SSN) was initially determined by the procedure of issuing numbers in groups of 10,000 to post offices for assignment on behalf of the Social Security Board's Bureau of Old-Age Benefits. The group numbers range from 01 to 99 (00 is not used), but for administrative reasons, they are not assigned consecutively. Within each area number allocated to a state, the sequence of group number assignments begins with the odd-numbered group numbers from 01 to 09, followed by even group numbers 10 through 98, then even numbers 02 through 08, and finally odd numbers 11 through 99.⁴

Serial Number

The last four digits of the SSN are the serial number. The serial number represents a straight numerical series of numbers from 0001–9999 within each group. Serial number 0000 is not assigned.

Designing the Social Security Card

Even at the inception of the program, the Social Security Board understood that individuals would need to have a "token" that would provide a record of the number that had been assigned to them. This token would help employers accurately report an individual's earnings under the program.

The Board first considered a small card similar to a credit union or trade union card, but some objected that it was too flimsy. Alternatively, a ¾ x 2⅞ inch metal card was proposed by a manufacturer of such cards. It was estimated that it would have taken 250 tons of metal for initial registration. The arguments in favor of the metal card were its permanence, accuracy (records could be imprinted from the embossed token), and economy (because of the imprinting capability). Still, in early June 1936, the Board decided to use a small paper card (McKinley and Frase 1970, 327 and 329).

In October 1936, the Social Security Board selected a design submitted by Frederick E. Happel, an artist and photo engraver from Albany, NY, for the original Social Security card, for which Happel was paid $60.⁵ The Board placed an initial order for 26 million cards. In late 1937, a second version was adopted, and a version just for replacement cards was adopted in 1938 (SSA 1990, 1). Since 1976, the design of original and replacement Social Security cards has been the same. In all, over 50 designs have been used from 1936 to 2008. All versions remain valid since it would be cost-prohibitive to replace all cards previously issued.

Over time, as the use of the SSN expanded for other purposes, SSA recognized that changes were necessary to protect the integrity of the card. SSA has taken measures to prevent counterfeiting of the card, and a counterfeit-resistant version is now used for both original and replacement cards. Steps taken by SSA to improve the card are detailed later.

Deciding on Application Data

There was also considerable discussion in 1936 about the types of information to collect as part of the registration. Generally, SSA collected the information needed to uniquely identify and accurately report an individual's earnings covered under the new Social Security program. Race was considered a necessary piece of information for actuarial purposes because of differences in life expectancy among different races. However, the Board decided to use the term "color" rather than race on the original Form SS-5 application for an SSN (McKinley and Frase 1970, 325–326).

The original 1936 version of the SS-5 requested the following information:

• Employee name
• Employee address
• Name of current employer
• Employer address
• Age of employee
• Date of birth
• Place of birth
• Sex
• Color
• U.S. Employment Service (USES) registration card number, if applicable
• Date and place of previously completed an SS-5, if applicable
• Completion date for current SS-5
• Signature (SSA 1990)

Registering the Nation's Employers and Employees

Although issuing SSNs is still a large workload for SSA, one rarely thinks about the major undertaking it was to register workers for the first SSNs. Initial estimates were that 22 million SSNs would be issued immediately, with 50 million ultimately to be issued (McKinley and Frase 1970, 15). In fact, 35 million SSNs were issued in the first 8 months of the registration effort. The Social Security Board estimated it would also need to assign identifying numbers to 3.5 million employers during this same time (McKinley and Frase 1970, 309).

Assigning responsibility for the vast registration process was a real problem. Debate shifted back and forth over whether the Board's Bureau of Old-Age Benefits could handle the work. The Board first approached the USES about assuming the registration workload, but in early May 1936 USES declined because President Roosevelt was hoping for an upswing in industrial production that autumn, and USES wanted its personnel to concentrate on its job placement service. The Census Bureau also declined, citing legal restrictions on the disclosure of its information to other agencies and confidentiality promises to the public that census information would be used for statistical purposes only (McKinley and Frase 1970, 338–339).

In June 1936, the Social Security Board decided that its Bureau of Old-Age Benefits would handle the registration and that the registration process would begin after the November 3, 1936, presidential election (McKinley and Frase 1970, 29). In May, the executive committee of the interdepartmental committee on enumeration had recommended that the Bureau could handle the registration by setting up 202 field offices and hiring 12,000 to 16,000 employees. The Board estimated that these 202 field offices would cover approximately 67 percent of registrants. On July 17, 1936, the Social Security Board's regional directors were told that 600 Bureau field offices would be open by November, that SSN assignment would begin about November 15, and that registration would be completed within 60 days. Also in July, the Board talked to the Post Office Department about assigning post office personnel to assist in cities where the Board would not yet have field offices to handle the registration (McKinley and Frase 1970, 342–347).

However, difficulties in recruiting personnel and setting up offices would make it impossible for the Bureau to handle the workload. As of September 30, 1936, Bureau of Old-Age Benefits employees numbered only 164 (Corson 1938, 6). Fortunately, the Board was able to enlist the Post Office Department to issue SSNs, signing an agreement on September 25, 1936. The Post Office Department had 45,000 facilities and over 350,000 employees at that time (Wyatt and Wandel 1937, 52).

The Social Security Board also enlisted the Treasury Department to assure employer cooperation. Final Treasury regulations, published in the Federal Register on November 6, 1936, required employers to file Form SS-4 (employer's application for an EIN) with the post office not later than November 21, 1936, and employees to file Form SS-5 (employee's application for an SSN) not later than December 5 (McKinley and Frase 1970, 15 and 360). However, delays in getting registration started made these deadlines moot.

The Social Security Board's Informational Service, established in January 1936, prepared a publicity campaign at midyear to encourage employers and workers to complete the application forms, but did not plan to distribute the material until after the November 3 election. However, the Board accelerated the publicity release in response to a September effort to discredit the program launched by Alf N. Landon, the Republican candidate for president. Also that year, many employers, in conjunction with Landon and the Republican Party, began stuffing payroll envelopes with leaflets against the Social Security Act and the required deductions from employee wages. The Social Security Board was so alarmed that the Chairman, John G. Winant, resigned in order to campaign in defense of the Social Security Act. In addition, in October 1936 the Board released a film called "We the People and Social Security" along with a 4-page pamphlet entitled "Security in Your Old Age."⁶ It is estimated that some 4 million people saw the film and nearly 8 million of the pamphlets were distributed by Election Day (McKinley and Frase 1970, 357–358).

On November 6, the campaign to encourage employers and employees to register began. A series of press releases outlined the procedure for assigning SSNs and carried sample Forms SS-4 and SS-5, as well as a Social Security card specimen. The campaign included three releases on old-age benefits in 24 languages distributed to the country's foreign language press. The Associated Press, the United Press, the Hearst chain, and many individual papers ran a series of articles on old-age benefits and registration for weeks at a time. During the November and December initial registration period, there were also 12 nationwide radio broadcasts by well-known individuals and a host of local broadcasts arranged by the 56 skeletal field offices then in place. Over 3 million posters⁷ were distributed, 50 million more pamphlets were dispersed, and three additional newsreel trailers were shown to some 42 million people (McKinley and Frase 1970, 364–366).

The registration process was largely directed by the local postmasters. The first task for the postmen was to make up lists of employers on their routes. Their effort resulted in a list of 2.4 million employers (McKinley and Frase 1970, 344–345 and 368).

Beginning November 16, 1936, the post offices sent Form SS-4s to employers based on the lists they had compiled earlier that month. Along with information about the business establishment, employers were asked for the number of workers they employed. The mail carriers collected the completed SS-4s a week or two later. Based on SS-4 information, the post offices delivered Form SS-5s to the employers the following week for distribution to employees (McKinley and Frase 1970, 368).

Employees were permitted to return the completed SS-5 application either to the employer, to any labor organization of which the employee was a member, to the letter carrier, to the post office by hand, or to the post office via mail (Wyatt and Wandel 1937, 54). This last possibility caused another round of negotiations between the Social Security Board and the Post Office Department about whether "return penalty privilege" (requiring no postage) applied. Postal regulations stated that this privilege could only be used if an individual was not required by law to submit the information. The Board argued that no postage was required as there was no law requiring employees to obtain an SSN. At the same time, however, the Board was requesting the Treasury Department to issue regulations mandating employees to obtain account numbers. In the end, the Board got it both ways—the Post Office Department agreed to accept returned SS-5s without postage on October 8, 1936, and the Treasury Department issued the regulations making the SSN mandatory on November 6, 1936 (McKinley and Frase 1970, 351–352 and 360).

Even at this early time, the public was concerned about privacy and confidentiality issues. Many employees were anxious to know how the information on the SS-5 would be used. The Social Security Board issued releases at various times assuring the public that the information on the application would be kept confidential, with access limited to government employees for whom job duties under the Social Security Act required it (Wyatt and Wandel 1937, 57). In June 1937, the Social Security Board would issue its very first regulation, formalizing its pledge of confidentiality for information collected and maintained.

The work of the Social Security Board to reassure the public was complicated by the actions of some employers, who circulated additional forms along with the SS-5. These extra forms requested information such as nationality, religion, education, and union affiliation. On February 26, 1937, the Board issued a press release warning employers against distributing unauthorized questionnaires that appeared to be required by the Social Security Board (Wyatt and Wandel 1937, 57).

Of the 45,000 post offices then in existence, 1,017 first-class offices were designated as "typing centers" to assign the SSNs, along with 57 "central accounting" post offices to assign SSNs for the second, third, and fourth class post offices within their area (McKinley and Frase 1970, 368). The Social Security Board supplied these centers with Office Record Form OA-702, in blocks of 1,000, with the account number preprinted. For each registrant, postal employees typed the information from the SS-5 onto the prenumbered OA-702 in duplicate. Each OA-702 had a detachable portion on which the employee's name was typed and then returned to the employee—the Social Security card. The post office mailed the completed Social Security cards to the employer, unless the employee had taken the SS-5 to the post office in person and waited for the typed card. Each completed card was accompanied by an informational circular briefly explaining the provisions of Title II (old-age benefits) and Title XVIII (the Social Security tax) of the Social Security Act (Wyatt and Wandel 1937, 54 and 58).

The post offices sent the completed SS-5 and the corresponding OA-702 forms in blocks of 500 to the Bureau of Old-Age Benefit's Records Office in Baltimore, where the SSN master files were to be kept. The post offices (and later the Bureau's field offices) kept carbon copies of the OA-702 to use should an individual request a replacement card (Wyatt and Wandel 1937, 58).

The publicity campaign and the Post Office Department's efforts resulted in over 22 million completed applications as of December 22, 1936, 28 days after the initial distribution of employee applications (Wyatt and Wandel 1937, 62). During the first 4 months of the registration campaign, nearly 26 million SSNs and more than 2.6 million EINs were assigned (Corson 1938, 3).

In September 1936, when the Post Office Department signed the agreement to handle the initial registration, the Board had planned for the Bureau of Old-Age Benefits to have 554 field offices set up to take over the enumeration workload in January 1937. However, hiring field staff had to await the results of the "examination for administrative officer" (civil service test) given in August 1936. The resulting register was not made available to the Board until December 1936 (McKinley and Frase 1970, 129). So in November 1936 the Board instead assigned headquarters staff to 56 Bureau field offices, covering all but one of the cities where the Post Office Department had set up its "central accounting" offices. These 56 Bureau offices primarily answered questions and directed applicants to the post offices (McKinley and Frase 1970, 34–35 and 369). The Board twice had to ask the Post Office to extend its handling of the SSN applications, first through March 1937 and then through June 1937, before the Bureau of Old-Age Benefits could take over.

Effective July 1937, Bureau field offices, still numbering only 175 with 1,702 total employees, took over the enumeration workload from the post offices (Zwintscher 1952, 90; SSA 1965, 25). By that time, some 35 million SSNs had been issued at a cost of $5.7 million (SSA 1990, 1; McKinley and Frase 1970, 372).

And still the job was not finished. In July 1937 alone, Bureau field offices issued some 1.9 million additional SSNs (McKinley and Frase 1970, 368–373). Even with field office employees working evenings and Saturdays and with "managers and assistant managers, anyone who was available, pounding away at typewriters," the Bureau had to set up additional typing centers in its 12 regional offices to help with the workload (SSA 1985, 10; SSA 1965, 32; SSA 1952).

Not all U.S. workers obtained SSNs during the initial registration period. This was because the original Social Security Act had excluded some types of employment from coverage, such as agricultural workers, domestic servants, casual labor, maritime workers, government employees, and the employees of philanthropic, educational, and similar institutions. The self-employed were also excluded from coverage. Seventy years ago, these exempt workers comprised about 40 percent of the working population. These groups were not covered primarily because of the administrative difficulty in collecting taxes and obtaining accurate wage reports (Department of Treasury 1947, 1; DeWitt, Béland, and Berkowitz 2008, 4).

Initially, only employees working in covered employment and aged 64 or younger were eligible to obtain an SSN. However, almost from the start, state unemployment compensation agencies began using SSNs to identify workers, and some employers tended to prefer hiring individuals who already possessed an SSN (Social Security Board 1938, 53). So, after a few months the Bureau began issuing SSNs to anyone who applied.

For over 20 years, Bureau field offices assigned SSNs, using blocks of prenumbered Social Security cards furnished to each office. Office staff simply typed the number holder's name on one of the prenumbered cards. For replacement cards, field office staff manually typed the SSN and name on a blank card. In 1961, issuance of original SSNs was centralized in Baltimore, but local offices continued to issue replacement cards. In March 1972, SSA began assigning SSNs and issuing cards exclusively from Baltimore via a computer-based system.⁸ It was also in 1972 that all applicants for federal benefits were required to have their own SSN.

Maintaining the SSN Records

Space to handle the SS-5 application forms was found on three floors of the Candler Building, a large warehouse converted from a Coca-Cola bottling factory on the harbor in downtown Baltimore (SSA 1961). Here the Bureau installed a "great battery" of International Business Machines (IBM) equipment⁹ and deployed over 2,300 machine workers and checkers by December 9, 1936, to handle the applications as quickly as they came in (McKinley and Frase 1970, 33 and 364). At this time, a hiring "apportionment" was in effect, which meant the Bureau could only recruit a certain proportion of employees from each state. As a result, employees came from all parts of the country. It was thought that the central operation in the Candler building was temporary, and that the work would soon be dispersed to the 12 regions, so recruitment from distant states was logical (SSA 1952). However, actual operations would reveal that decentralization was not really feasible.

The Baltimore Records Office used a nine-step process to create a permanent master record and to establish an earnings record for each individual.¹⁰ One hundred applications and office record cards, numbered consecutively, were sent through each operation together with a control unit of nine cards (one for each step). The appropriate control card was removed at the end of a step and sent to a control file to track the status of each block (McKinley and Frase 1970, 375).

When the Records Office received the Form SS-5 and the accompanying OA-702 from the local offices, different clerks working independently converted the two sets of information into numerical codes that could be transferred to punch cards.

The first group of employees keyed information from the SS-5 into a master punch card for each individual. A tabulating machine used this master punch card to set up a numerical register of accounts stored in huge loose-leaf books. These volumes contained the SSN, name, and date of birth of each number holder. Each page contained 100 SSNs in numerical order. From these volumes, employees could learn the name and identifying information of an SSN's owner in a fraction of the time that would be required to locate the master punch card (Wyatt and Wandel 1937, 120–121). The master punch card was also used in the earnings-posting operation to establish an earnings ledger for each individual.

A second group of employees independently keyed the same information coded from the OA-702 to create an actuarial punch card (Fay and Wasserman 1938, 25). The actuarial punch card was created for actuarial and statistical purposes and was also used to set up the "visible index." Later known as the "National Employee Index Flexoline File" or just "Flexoline," the visual index consisted of strips of thin bamboo, 3/16 of an inch wide by 9 inches long—one for each SSN issued—set in a steel panel. The strips were inserted into the frame one by one, with some employees filing as many as 300 strips an hour. Each strip began with a 3-digit entry based on the Russell Soundex System (in which all surnames having the same basic consonants are grouped together), followed by the individual's surname, given name, middle initial, and SSN. The strips were mechanically prepared from the actuarial punch card and manually posted on the panel, sorted by the first letter of the surname and within each letter by phonetic code, then in each code group by the first seven letters of the first name, middle initial, year and month of birth, and SSN. Up to 1,600 panels were then hung on each rack (Staruch 1978, 29). The primary function of the visible index was to aid in the location of accounts when only the name of the owner and not the SSN was known (Wyatt and Wandel 1937, 121). For instance, employees referred to the index when a worker who had lost his or her card and did not know the SSN applied for a duplicate (SSA 1964 and Staruch 1978, 29). Reportedly, experienced clerks were able to find any name and its corresponding account number in less than 60 seconds (Fay and Wasserman 1938, 25).

In addition, the SS-5s were filmed on 16 millimeter, noninflammable film strips. In June 1938, officials bragged "This film is so compact that the entire file of 40 million photographed SS-5s is stored in 10 ordinary letter-size file cabinets" (Fay and Wasserman 1938, 25).

In all, eight separate files were maintained:

• The SS-5 applications, sorted in numerical order.
• Photographs on 16 millimeter film strips of the SS-5s, in numerical order.
• The master punch cards, in numerical order.
• The numerical register, in large loose-leaf books.
• Ledger sheets for maintaining earnings records, in numerical order.
• The OA-702 Office Records, in alphabetical order.
• The actuarial punch cards, in phonetic code surname order.
• The visible index, in phonetic code surname order (Fay and Wasserman 1938, 25–26).

By 1958, the Flexoline (visible index) contained 160 million strips in 750 steel A-frame stands, and SSA was adding an average of 7 million new strips each year. In August 1958, SSA began converting the Flexoline index to microfilm and began capturing new SSN records on magnetic tape, using a special machine to then transcribe the code directly from magnetic tape into a readable microfilm record (Staruch 1978, 29–30). By 1964, the 200 million names in the National Employee Index were contained on 2,005 reels of magnetic tape that Bureau employees accessed by means of high-speed microfilm readers (SSA 1964).¹¹

In 1972, SSA created an electronic file, the Numerical Index File or Numident, to house the numerically-ordered master file of all assigned SSNs. In 1973, SSA began converting its legacy SS-5 records to the Numident electronic database, completing the conversion in 1979. There is one Numident record for each SSN ever assigned.

SSA makes changes in Numident SS-5 data only upon receipt of updated information from the SSN holder. Changes in the Numident result in the addition of a new entry or iteration to the Numident record for the individual; information is never overlaid on a previous SSN Numident entry.¹² Most changes are initiated when an SSN holder completes an SS-5 requesting a replacement card or a change in the name, sex, or date of birth information on the Numident. Additionally, SSA employees may take action to change identifying information on the Numident for a person while taking a claim or processing postentitlement events. Each Numident record can contain up to 300 Numident entries (iterations) representing an addition or change to the Numident information for a person. About half of Numident records have multiple entries.

Until recently SSA also maintained a separate SSN master file indexed by cardholder name. The Alpha Index File or Alphident enabled SSA employees to search by name if the number was unknown. In the process of modernizing SSA's master files, this file was converted to an IBM DB2 relational database linked to the Numident file. This database provides the same basic functionality as the Alphident. Like the Flexoline, the DB2 uses the Russell Soundex Coding System to group all surnames that have the same basic consonant sounds. When an individual's identifying information is available, an SSA employee can attempt to locate the SSN using a key based on the Soundex version of the last name, plus the first 4 characters of the first name, plus the century, year, and month of birth. SSA has designated this database a sensitive file and access is restricted.

Handling SSN Assignment Problems

From the beginning, the process of assigning SSNs included quality checks. SSA employees had to account for every number and explain any missing serial numbers fully. Also, the SS-5s and the OA-702s were coded separately by different clerks and were later compared as a quality check (Fay and Wasserman 1938, 24).

Still, as one might expect, an undertaking as enormous as enumerating 35 million workers in one concentrated effort was bound to encounter some problems. Many individuals received multiple SSNs. Some people were under the impression that the more SSNs they received, the better. Others thought they needed a new SSN for each new job. Workers sometimes lost their original number and applied for a new one. Also, a great many unemployed and WPA employees applied for SSNs both during the initial registration and again through WPA or private employment registration. Sample studies in 1937 or early 1938 indicated that duplicate account numbers had been issued to not more than 3 or 4 percent of the applicants (Corson 1938, 4).

In 1938, a wallet manufacturer in Lockport, New York, the E.H. Ferree Company, decided to promote its product by showing how a Social Security card would fit into the wallet. The company vice president thought it would be clever to use a sample card with his secretary's actual SSN. The wallet was sold at Woolworth's¹³ and many other large department stores, and the SSN was widely distributed. Many purchasers adopted the SSN as their own—5,755 people were using it in the peak year 1943, and 12 were still using it as late as 1977. In all, SSA received 40,000 incorrect earnings reports under this SSN, which had to be reassigned laboriously to proper SSNs. SSA voided the much-used number and issued a new SSN to the secretary (SSA n.d. c).

About a dozen similar cases of individuals adopting a made-up SSN shown on a facsimile card have occurred. In one case, the Social Security Board itself issued a pamphlet with the made-up number 219-09-9999 that was adopted by an individual (SSA n.d. c).

Also, prior to 1961 SSA field offices issued new SSNs. Only a fraction of these SSN assignments were screened at the central office for a previously assigned SSN, and then only manually (Long 1993, 84). Thus, issuing duplicate SSNs was possible. Beginning in 1961, the central office in Baltimore issued all new SSNs, but it was not until 1970 that an electronic method of checking for previously issued SSNs (called "EVAN" for "electronic verification of alleged numbers") was devised (SSA 1990, 4). Today, automated systems with sophisticated matching routines screen for previously issued SSNs.

SSA has since introduced more rigorous verification procedures. On April 15, 1974, SSA implemented evidence requirements (age, identity, and citizenship/alien status) for applicants for an original SSN who are foreign-born, or are U.S.-born and age 18 or older. Then, on May 15, 1978, SSA began requiring evidence of age, identity, and citizenship/alien status from all applicants for original SSNs, and evidence of identity for replacement Social Security cards. In addition, all foreign-born applicants for replacement cards were required to submit evidence of citizenship/alien status.

Also, in 1979 SSA created an electronic file called MULTX from a set of punch cards identifying multiple SSNs that was maintained by SSA's Office of Earnings Operations. As of December 2007, SSA had identified and cross-referenced in the MULTX file over 4.7 million individuals with multiple SSNs, about 93 percent of whom have only two SSNs. Generally, those with multiple SSNs are the "very old" on the Numident; a study conducted in 2002 showed a weighted average age of 82.9 (SSA 2002). The requirement for proof of age and identity for SSN applicants beginning in 1974 combined with the implementation of an automated SSN screening system in 1984 have significantly reduced the multiple-SSN problems.

Under a few rare circumstances, SSA may legitimately issue a new SSN to a person with a prior SSN. The conditions are highly restrictive. SSA will assign a new SSN to a victim of harassment, abuse, or life endangerment if the individual provides evidence to substantiate the allegations. In addition, SSA may assign a new SSN to an individual who is a victim of SSN misuse, which means that the number has been used with criminal or harmful intent and the individual has been subjected to economic or personal hardship. Third party evidence is necessary for SSA to substantiate an individual's allegation of SSN misuse. However, an individual should consider changing his or her SSN only as a last resort because getting a new SSN may adversely impact one's ability to interact with federal agencies, state agencies, and employers, as all of the individual's records will be under the former SSN.

Applying for an SSN Today

Just as it was in 1936, today a person must complete an application to obtain an original or replacement SSN or to change the information in SSA's Numident records. There are a number of ways to initiate the application process.

The paper form a person completes to apply for an original SSN or a replacement card or to make changes to SSA's Numident record is still the SS-5. The SS-5 application is available online¹⁴ or in any SSA field office. The application and required evidence can be taken or mailed to any Social Security office for processing. An in-person interview is required if the applicant is age 12 or older and is applying for an original SSN. The Veterans Affairs Regional Office (VARO) in Manila also accepts SS-5 applications for an original SSN or a replacement card, as do all U.S. Foreign Service posts and all military posts outside the United States. SSA employees key the SS-5 application data and evidence into the SSA computer system, which uses the information to create or update the Numident. The signed SS-5 application is retained for a short period in the field office, and then is sent to a records center in Pennsylvania for microfilming. Once microfilmed, the original SS-5 is destroyed.¹⁵

In August 1987, SSA began a three-state pilot of the "Enumeration at Birth" (EAB) process in which the parent of a newborn can request an SSN as part of the state's birth registration process. Additional states began to participate in EAB in July 1988. By the end of 1991, 45 states, the District of Columbia, Puerto Rico, and New York City had signed agreements (Long 1993, 83). Today, over 90 percent of parents use the EAB process offered in all 50 states plus Puerto Rico and the District of Columbia. SSA receives nearly three-quarters of original SSN applications through the EAB process and issues over 4 million SSNs via EAB each year (SSA 2006). No microfilm SS-5 exists for a record created through the EAB process.

Beginning in 2002, SSA began another pilot program referred to as Enumeration at Entry (EaE) that allows noncitizens admitted for permanent residence to obtain SSNs and Social Security cards based on data collected as part of the immigration process. This pilot was expanded worldwide in early 2003. EaE is a joint effort involving the Department of State (DoS), DHS, and SSA. Under EaE, a person aged 18 or older can apply for both an immigrant visa and an SSN at a DoS office in his or her home country. If the visa is granted, then DoS transmits the identifying data from the person's visa/SSN application to DHS. If and when the person is physically admitted to the United States, DHS updates certain data, if necessary, and sends it to SSA for the SSN to be assigned and the card to be issued. All noncitizens enumerated through EaE receive an SSN in the special area number series 729 through 733. As of January 20, 2009, SSA had issued 429,959 original and 114,714 replacement SSNs through the EaE process. SSA is currently working with DoS and DHS on expanding the EaE process to additional noncitizens.

Also in 2002, SSA began to open offices dedicated entirely to handling Social Security number business. The first Social Security Card Center (SSCC) opened in Brooklyn, NY, in November 2002. Six more SSCCs have since opened: Las Vegas, NV, in April 2005; Jamaica, NY, in July 2006; Downtown and North Phoenix, AZ, in October 2007; Orlando, FL, in March 2008; and Sacramento, CA, in November 2008. Generally, any individuals who live in the service area of a Card Center and need an original or replacement card must visit the Card Center rather than their local field office.

The Intelligence Reform and Terrorism Prevention Act (IRTPA) of 2004 (Public Law (P.L.) 108-458) placed limits on the number of replacement Social Security cards an individual can receive. Beginning with cards issued on or after December 17, 2005, individuals may only receive three Social Security cards per year and 10 in a lifetime, with certain exceptions, such as correcting errors or name changes.

The information currently requested on the SS-5 is:

• Name to be shown on the card
• Full name at birth, if different
• Other names used
• Mailing address
• Citizenship or alien status
• Sex
• Race/ethnic description (SSA does not receive this information under EAB)
• Date of birth
• Place of birth
• Mother's name at birth
• Mother's SSN (SSA collects this information for the Internal Revenue Service (IRS) on an original application for a child under age 18. SSA does not retain these data.)
• Fathers' name
• Father's SSN (SSA collects this information for IRS on an original application for a child under age 18. SSA does not retain these data.)
• Whether applicant ever filed for an SSN before
• Prior SSNs assigned
• Name on most recent Social Security card
• Different date of birth if used on an earlier SSN application.
• Date application completed
• Phone number
• Signature
• Applicant's relationship to the number holder

Evidentiary Requirements

At the inception of the program, all SSNs were assigned and cards issued based solely on information provided by the applicant. However, in the 1970s, SSA began requiring proof of age, identity, and citizenship.

SSA has instituted numerous evidentiary requirements to further safeguard and preserve the integrity of the SSN and to ensure assignment of SSNs and issuance of cards only to eligible individuals. Exhibit 1 shows the effective dates of changes in policy on evidentiary requirements.

Expanding Uses of the SSN

The original purpose of the SSN was to enable the Social Security Board to maintain accurate records of the earnings of individuals who worked in jobs covered under the Social Security program. The card was never intended to serve as a personal identification document—that is, it does not establish that the person presenting the card is actually the person whose name and SSN appear on the card. Although SSA has made the card counterfeit-resistant, the card does not contain information that would allow it to be used as proof of identity. However, the simplicity and efficiency of using a unique number that most people already possess has encouraged widespread use of the SSN by both government agencies and private enterprises, especially as they have adapted their recordkeeping and business systems to automated data processing. Use of the SSN as a convenient means of identifying people in large systems of records has increased over the years and its expanded use appears to be an enduring trend. Generally, there are no restrictions in federal law precluding the use of the SSN by the private sector, so businesses may ask individuals for an SSN whenever they wish (Streckewald 2006).

The expansion of SSN use began in 1943 with Executive Order (EO) 9397 requiring federal agencies to use the SSN for the purpose of identifying individuals in any new record systems. Although there was considerable delay in other agencies adopting its use, the coming of the computer age in the 1960s led government agencies and private industry alike to find many uses for the SSN.

In 1971, an SSA task force studied issues raised by nonprogram use of the SSN and proposed that SSA take a "cautious and conservative" position and do nothing to promote its use as an identifier. In 1973, a report of the Department of Health, Education, and Welfare (now Health and Human Services) concluded that the adoption of a national identifier was not desirable, and that the SSN was not suitable for such a purpose (SSA 1997). Nevertheless, Congressional legislation and federal agency regulations require the collection of SSNs for myriad purposes, as detailed in Exhibit 2.

With the many purposes legally requiring an SSN, the need for a U.S. resident to possess one has become nearly universal. The universality of SSN ownership has in turn led to the SSN's adoption by private industry as a unique identifier.

Unfortunately, this universality has led to abuse of the SSN. Most notoriously, the SSN is a key piece of information used to commit identity theft. According to the Federal Trade Commission (FTC), "Identity theft occurs when someone uses your personally identifying information, like your name, Social Security number, or credit card number, without your permission, to commit fraud or other crimes." The FTC estimates that as many as 9 million Americans have their identities stolen each year (FTC n.d.). Identity theft has reached such proportions that President George W. Bush issued Executive Order 13402 on May 10, 2006, establishing the President's Identity Theft Task Force. The task force reported:

The simplicity and efficiency of using a seemingly unique number that most people already possessed encouraged widespread use of the SSN as an identifier by both government agencies and private enterprises, especially as they adapted their record-keeping and business systems to automated data processing. The use of SSNs is now common in our society.

Employers must collect SSNs for tax reporting purposes. Doctors or hospitals may need them to facilitate Medicare reimbursement. SSNs also are used in internal systems to sort and track information about individuals, and in some cases are displayed on identification cards. In 2004, an estimated 42 million Medicare cards displayed the entire SSN, as did approximately 8 million Department of Defense insurance cards. In addition, although the Veterans Health Administration (VHA) discontinued the issuance of Veterans Identification Cards that display SSNs in March 2004, and has issued new cards that do not display SSNs, the VHA estimates that between 3 million and 4 million previously issued cards containing SSNs remain in circulation with veterans receiving VA health care services. Some universities still use the SSN as the students' identification number for a range of purposes, from administering loans to tracking grades, and may place it on students' identification cards, although usage for these purposes is declining.

SSNs also are widely available in public records held by federal agencies, states, local jurisdictions, and courts. As of 2004, 41 states and the District of Columbia, as well as 75 percent of U.S. counties, displayed SSNs in public records. Although the number and type of records in which SSNs are displayed vary greatly across states and counties, SSNs are most often found in court and property records (President's Identity Theft Task Force 2007, 23–24).

Verifying SSNs

Because individuals sometimes use SSNs that do not belong to them, either through error or deliberately, it is important to ensure that an SSN matches SSA records before accepting it.

Today, SSA electronically verifies that an SSN and the name associated with it match those in SSA's records before issuing a replacement Social Security card, posting a wage item to the Master Earnings File, or establishing a claims record. Also, when disclosure laws allow, many federal and state agencies use an SSA verification system to verify SSNs. Registered private employers can also verify a worker's SSN. In addition, SSA receives requests for SSN verification from third parties who have obtained the consent of the individuals involved.

However, SSNs were in use for many years before electronic verification was in place. During the 1950s, SSA initiated a manual screening routine of the microfilm file to search for a previously assigned number.

SSA's first electronic system was not developed until 1970, when electronic verification of alleged numbers (EVAN) was introduced for internal use at field offices with Advanced Records System (ARS) submission. In 1975, SSA's Bureau of Data Processing implemented the Full Registration and Identification System (FRIS) which expanded electronic screening capabilities and added electronic validation of SSNs (SSA 1990, 5). In 1983, SSA implemented the Enumeration Verification System (EVS) for verifying batches of SSNs; EVS employed a series of verification routines that are still in use. In 1984, SSA created the Automated Enumeration Screening Process (AESP) to run every application for an original or replacement card through the Alphident (since converted to a database linked to the Numident file) to determine if the data on the incoming record match one or more existing records using a complex scoring system. If a potential match is indicated between an existing record and an application for an original SSN, the field office is alerted to resolve the matter. If no previously established record is found, an original SSN is issued to the applicant (SSA 2003, 8–12).

In more recent years, SSA has developed a number of SSN verification systems for internal and external use. Having multiple systems allows SSA to customize the input and output. Those for external entities generally indicate whether the data submitted match SSA records, and whether the SSN holder has died, but will not disclose additional information to the requestor. A few examples are described below.

The Consent Based SSN Verification Service (CBSV) is available to enrolled private companies and federal, state, and local agencies to verify that the submitted name and SSN match SSA records. The recent consent of the SSN holder to release the information is required.¹⁶

When a person lacking a valid driver's license registers to vote, the individual can provide his or her name, date of birth, and the last four digits of his or her SSN instead. The state then submits this information to SSA's Help America Vote Verification (HAVV) system to verify that the submitted data match SSA records.

State Departments of Motor Vehicles use the Social Security OnLine Verification (SSOLV) system to verify names and SSNs for the issuance of new and renewal driver's licenses and identity cards.

The SSN Verification Service (SSNVS) is a free Internet-based system that can be used by registered employers for SSN verification prior to wage reporting. Real-time service is provided for 10 requests or less, and overnight processing is provided for up to 250,000 SSNs.¹⁷

The E-Verify program (previously known as the Basic Pilot/Employment Eligibility Verification System), administered by DHS with SSA's support, can be used by employers to verify the SSN and confirm employment authorization under immigration law for newly hired employees. The employer enters the name, SSN, date of birth, and alleged citizenship/alien status from DHS Form I-9 Employment Eligibility Verification into the E-Verify system. E-Verify automatically matches this information against SSA's Numident, as well as DHS immigration records if the hire is a noncitizen. The employer receives an electronic response indicating either that employment is authorized or that the data do not match the information in SSA's or DHS' records. Use of the verification program is voluntary in most parts of the United States, but 13 states require certain employers to use it for new hires. The U.S. Citizenship and Immigration Services website indicates that 109,211 employers, representing close to 434,000 worksites, were registered to use E-Verify as of February 2009.

A proposed amendment to the 2009 Omnibus Appropriations Act to extend authorization for the E-Verify program for 5 years was tabled; the bill instead extended authorization through September 30, 2009.

Enhancing the Social Security Card

In addition to developing verification systems to allow authorized users to determine if SSN information matches SSA data, SSA has also taken steps to help guard against fraudulent Social Security cards. Because an SSN is needed for work and has been adopted for many other uses, a market for counterfeit Social Security cards has developed.

SSA has taken and continues to take steps to strengthen the integrity of the Social Security card and guard against its misuse. One of the first steps was to distinguish whether cards were valid for work purposes. In 1974, SSA began assigning SSNs for nonwork purposes when such use of an SSN was authorized by law. Initially, the nonwork cards looked the same as cards issued to citizens and aliens authorized to work. In May 1982, SSA began annotating cards issued for nonwork purposes with the legend "not valid for employment."

The 1981 Omnibus Reconciliation Act (P.L. 97-123) added alteration and forgery of a Social Security card to the list of prohibited acts and also increased the penalties for such acts. In 1983, section 205(c)(2)(G) of the Social Security Act was amended to require that the "social security card shall be made of bank-note paper and (to the maximum extent practicable) shall be a card which cannot be counterfeited" (P.L. 98-21). SSA worked with the Bureau of Engraving and Printing, the Secret Service, and the Federal Bureau of Investigation (FBI) to design a card that met these requirements. Changes were made to the card stock to make any attempt to erase or remove data easily detectable, and a form of printing with a raised effect that is difficult to replicate was used. Other features not obvious to the naked eye were also added.

In 1988, to prevent photocopy counterfeits, a security feature that displays as "void" when photocopied was added. Also in 1988, legislation increased the monetary penalties for SSN violations. In September 1992, SSA began to annotate Social Security cards for aliens with temporary work authorization "valid for work only with INS [now DHS] authorization."

Immigration and welfare reform legislation enacted in August 1996 (P.L. 104-208 and P.L. 104-193, respectively) directed SSA to develop a prototype of a counterfeit-resistant Social Security card and report to Congress on the feasibility of using such cards. The prototype was to be made of a durable, tamper-resistant material such as plastic or polyester; was to employ technologies that provide security features, such as magnetic stripes, holograms, and integrated circuits; and was to provide individuals with reliable proof of citizenship or legal resident alien status.

SSA interpreted this provision as calling for consideration of techniques that could link the card to the assigned number holder, such as including identifying information about the card holder on the card itself, or adding the number holder's picture, fingerprint, biometric identifier, or some combination of such features in or on the card (Donnelly 1999).

In September 1997, SSA published its findings in the Report to Congress on Options for Enhancing the Social Security Card. The seven options that SSA developed included:

• Plastic card
• Card with the number holder's picture
• Card with a secure barcode data storage stripe
• Card with an optical memory storage stripe
• Card with magnetic stripe
• Card with magnetic stripe and the number holder's picture
• Microprocessor card with a magnetic stripe and a picture

SSA estimated that the cost at that time of issuing an enhanced card to almost 300 million card holders would range, depending on the type of card, from $5.1 billion to $10.5 billion. The estimates included the costs of verifying the identity of the applicant and establishing a system to collect a user fee (Donnelly 1999). SSA was not required to adopt any of the options. To some extent, passage of the REAL ID Act in 2005, which imposes standards on states for the issuance of driver's licenses and identification cards, diminished the need for SSA to develop a card for identification (SSA 2008).

In 2004, the IRTPA legislation required SSA, in consultation with DHS, to form an interagency task force to establish requirements for further improving the security of Social Security cards and numbers and to provide for implementation of those requirements. The task force was formed in January 2006 and included several other agencies, such as the FBI, the DoS, and the Government Printing Office. The task force issued its recommendations in May 2006. As a result, additional security features were added to the card in 2007. These include:

• Latent images that can be seen when the card is viewed at an angle.
• Color shifting ink (similar to that used on $20 bills) that changes colors when the card is viewed from different angles.
• A new production method that blends different color inks on the background of the card (colors flow from blue to aqua).
• Adding issuance date to the front of the card.

Other features added to the card are not apparent to the naked eye and for security reasons are not disclosed to the public.

Another change made to the Social Security card in 2007 was to put the first name and last name of the card holder on separate lines. This change was recommended by employer groups to help them distinguish the correct last name of an employee.

The expertise of counterfeiters and the widespread availability of state-of-the-art technology make it increasingly difficult to develop and maintain a document that cannot be counterfeited. SSA continues to evaluate new technology as it becomes available to determine if additional features should be included to make it more difficult to alter or counterfeit the card. In addition to the physical changes made to the Social Security card, SSA has taken many other steps to strengthen the integrity of the enumeration process by requiring evidence of age, citizenship, and identity, and by verifying this information, as noted in the section on evidentiary requirements.

Conclusion

The use of the SSN has expanded substantially since its inception in 1936. Created merely to keep track of the earnings history of U.S. workers for Social Security entitlement and benefit computation purposes, it has become a number assigned at birth and used by many government agencies to identify individuals and by private industry to track an individual's financial history.

That trend has begun to shift. As early as December 2004, IRTPA legislation prohibited states from displaying the SSN on driver's licenses or motor vehicle registrations. In 2007, the President's Identity Theft Task Force (2007, 3) included among its SSN recommendations that "federal agencies should reduce the unnecessary use of SSNs, the most valuable commodity for an identity thief."

On November 18, 2008, President George W. Bush issued EO 13478 rescinding the 1943 EO requiring all federal agencies to use the SSN as an identifier. Then in December, the FTC (2008) issued a plea to companies, schools, and other private entities to find better ways to authenticate identities than using the SSN. State and local entities have begun to delete SSNs on electronic versions of public records. Congress has also considered legislation that would require the Centers for Medicare and Medicaid Services to use an alternative to the SSN as the Medicare claim number. Even SSA, which created the SSN for its program use, has ceased to print the full SSN on some of its correspondence with beneficiaries (Lockhart 2002). The agency now advises individuals to keep their Social Security card in a safe place and not to carry it with them (SSA 2007a).

Federal survey-takers are also finding that as respondents have become more aware of the risk of identity theft, they are less willing to supply SSNs that are useful in linking the agency's survey data with administrative records from other agencies. For the Census Bureau's Survey of Income and Program Participation, the share of respondents who did not provide their SSN increased from 12 percent in 1996 to 35 percent in 2004. Likewise, the share of respondents who did not provide an SSN for the Census Bureau's Current Population Survey increased from 10 percent in 1994 to almost 23 percent in 2003 (Bates 2004, 4). The National Center for Health Statistics reports a similar problem.

Still, it is highly unlikely that use of the SSN as a unique identifier will cease entirely. In order to share data among government agencies or between commercial firms, a unique identifier to match records is critically important, and the SSN is the one unique tag that follows an individual throughout life. People may change their names and addresses throughout their lives, but their SSNs generally will remain the same. Of course, the SSN will also still be used for its original purpose—to track earnings in SSA records. The SSN is here to stay for the foreseeable future.

SSA will continue to assess its policies and procedures to further strengthen the integrity of the enumeration process to prevent SSN fraud and misuse, as well as to protect the important personal information with which it is entrusted. SSA is a member of the Identity Theft Task Force and will continue to work with other members to protect, to the maximum extent possible, the integrity of the SSN.