Mr. Chairman I am happy to have the chance to appear today before this committee on the complex question of the proper use of the social security account number as a national identifier. The Center for Immigration Studies, which I head, is a non-profit research and policy analysis group, one of whose objectives is to examine the means for humane and effective enforcement of our immi-gration laws in ways consistent with democratic values. Central to this undertaking is consideration of the appropriate use of non-intrusive secure identi-fication systems in support of such legislated objectives as enforcement of employer sanctions, the screening of non-entitled aliens from public assistance rolls, and federal prohibitions on ownership of firearms by illegal aliens and persons who have lost their US citizenship.
The appropriate balancer among the imperatives of security, efficiency and privacy of national identification systems is a question now facing other western industrial nations along with the United States. Recognizing the transnational character of this issue, the Center for Immigration Studies sponsored in November of 1990 an international conference of identification and privacy experts from thirteen industrial nations, including the Soviet Union and Japan. Discussions ranged over such issues as the appropriate purposes of and safeguards for national identification numbers; special purpose identification numbers such as those for tax, social insurance, voter registration and property ownership; national identification cards for citizens and aliens; and national population registers. The conference also addressed safeguards for the security and privacy of vital records
An important issue at the conference was the nature and extent of the threat of abuses of privacy from national ID numbers, or from special purpose numbers now broadly used, such as the US social security number and Canada's social insurance number (SIN), because of their facilitation of computer matching. There was general agreement that special use numbers covering broad populations become a valuable public administration resource that tends to invite use for purposes not originally intended. But representatives of the Western European continental democracies showed greater pragmatism than the US and other common-law countries (UK, Australia and Canada) in accepting the public administration advantages of such systems. Moreover, unlike the United States, nearly all Western European countries have the extra protection of centralized, semi-autonomous data protection agencies or privacy commissions to guard against abuses of ID numbers and related data, backed up by legislation based on international privacy standards.
There was considerable discussion of whether ID numbers do in fact facilitate the matching of data to a dangerous degree. The assumptions underlying that proposition were questioned. It was noted that the existence of ID numbers as identifiers is no longer a necessary condition for efficient matching of personal data; in any event, unauthorized matching is best discouraged by means other than the imposition of inefficient identifiers.
Evidence was presented showing that highly efficient matching is possible without ID numbers, using such identifiers as names and birth dates - an efficiency that steadily improves with advances in computer technology. The second questionable assumption - that technical self-limitation is an effective safeguard - is particularly applicable to the US experience. As Norwegian delegate Knut Selmer expressed it
The defense of democratic values must be sought elsewhere. The suppression of birth numbers or other technical means will only postpone the troubles; technical obstacles could always be circumvented. There must, in other words, be a constant awareness in political circles that data technology can endanger our democratic values. Furthermore, there must be potent public organs with the single task of supervising the use of computers in the administration, and sounding the bell if the development gives rise to anxiety (Selmer, 1990).
I share this view. The massive collections of data now held by Federal and state agencies and managed with the most sophisticated data processing techniques can be seen as a rich resource rather than as Leviathan, not as a threat but as an opportunity for highly efficient public administration and provision of services that improve our quality of life, our protection from crime and abuse, and even our privacy. Self-imposed inefficiencies in the management of such a resource as a protection against abuse is wasteful and ultimately self-defeating. An enlightened democratic society is capable of devising far better and more flexible safeguards.
IDENTIFICATION ANARCHY UNDERCUTS
IMMIGRATION CONTROL
It is hardly news that this country's diverse and decentralized non-system of personal identification is what the Reverend Theodore Hesburgh called anarchic, an open invitation to massive fraud and abuse and a serious impediment to effective law enforcement, including compliance with the immigration laws. Such a system also entails its own unique threats to individual privacy. The Federal Advisory Committee on False Identification (FACFI) amply documented this broad pattern of abuse in its 1976 report. While the situation has improved a little since the mid-1970s, Congress has since approved a series of laws that demand better identification for their effective enforcement or which increase the penalties on those engaging in abuses of documents. Congress in the 1984 Omnibus Crime Control Act required Federal agencies to move toward greater commonality in their own identification systems and directed the Executive Branch to submit recom-mendations for secure federal identification systems by 1987. Unfortunately the Executive has ignored this Congressional instruction.
No Federal enforcement effort has been more hampered by document fraud than sanctions against those who employ unauthorized aliens. Pervasive and easily available counterfeit documents purporting to establish work authorization are seriously under-mining the deterrent effect of the 1986 immigration reforms, eroding basic respect for the law in general. The resurgence of illegal immigration in 1990, as indicated by a sharp rise in apprehensions of illegal aliens after three years of decline, suggests that once- hesitant candidates for illegal entry into the American workplace are now convinced that employer sanctions are easily evaded.
No federal enforcement effort has been
more hampered by document fraud
than sanctions against those who
employ unauthorized aliens.
Looking at the diversity and unreliability of the crazy-quilt of documents authorized by the Immigration and Naturalization Service (INS) to prove identity and/or work eligibility, it's hard to see how enforcement of sanctions could be anything but confused - for employers and workers as well as enforcement officials. INS originally designated 17 different documents as acceptable for use. But a number of the acceptable documents are in themselves broad categories of diverse documents of varying quality and reliability, further taxing the employer's ability to judge authenticity among so many different permutations. Some examples
? Drivers licenses, issued with different formats, security standards and personal data by 50 states and five other jurisdictions.
? Baptismal records, issued without uniform standards or formats by nearly 100,000 parish churches that practice infant baptism.
? Social Security cards, with more than 15 versions in use, and only the most recent with any counterfeit-resistant features.
? Voter registration cards, issued in different formats by the 50 states, often by mail with no confirmation of identity or legal residence status.
? INS green cards, with three versions in use, two of them readily counterfeitable.
? A basic breeder document for other ID, the birth certificate is issued by some 7,000 different jurisdictions, with widely varying formats and security arrangements.
Besides undermining enforcement, this jumble of easily abused documents has placed otherwise well-intentioned employers in a dilemma many have responded by accepting suspect documents no questions asked; while others have unintentionally discriminated by insisting on specific documents of presumably greater reliability from job seekers.
OPTIONS FOR BETTER VERIFICATION
A SOCIAL SECURITY CALL-IN SYSTEM
So the identification arrangements set in place since the 1986 Immigration Reform and Control Act (IRCA) have clearly failed. Now is the time to move to more secure systems as that law envisaged. But while the responsible agencies have completed the mandated studies of the options for telephone-in systems, secure social security cards, and employer validation of the social security numbers of job candidates, there has been little movement toward actual choice and implementation of alternative systems in the nearly five years since IRCA.
...I believe that prudent use of the
social security number can greatly
contribute to current efforts to
make basic ID documents more fraud
proof and establish effective
verification systems...
INS has been stalled for nearly three years in getting the go-ahead for a pilot project to test a phone-in system of verification. The system to be tested is at best modest in scope and only partially responsive to broad enforcement needs. It will provide verification only to those job seekers who are - or who admit to being - foreign born and it relies on a data base keyed to the Alien ID number or A number, rather than the social security number.
Legislation failed in the last Congress to authorize a limited pilot project in a few states to examine enhanced security features, including biometric identifiers and personal data, that would make the drivers license or non-drivers state identification card a key document in worker verification.
Mr. Chairman, I believe that prudent use of the social security number can greatly contribute to current efforts to make basic ID documents more fraud proof and establish effective verification systems, without violating legitimate privacy concerns or loading major uncompensated admini-strative costs on the government.
Over the longer term, if the burdens and perplexities of the nation's seven million employers are to be eased, IRCA-related job discrimination banished, and unauthorized employment ended, it will be necessary to move to a universal call-in system (using phone, modem, or point of sale device) based on the social security number that would verify the 50 to 60 million hiring transactions that occur in the US economy each year. While such a system would entail added administrative costs, these would be offset by reduced costs of workplace investigations and other on-site enforcement measures that would be obviated.
The amount of individual data on nearly 250 million social security enrollees required for such a verification system should be limited to the minimum required to establish positive identification. The data base containing it should be administered separately from the main files of the Social Security Admini-stration. If a transaction number were provided to employers for each phone-in verification, the employer would have a record of compliance, ultimately making it possible to dispense with the burdensome I-9 verification form.
There would be other administrative benefits as well. Verification of job applicants against social security account data would aid in ridding the system of invalid or inactive numbers and in controlling use of numbers by more than one person or by imposters. Avoidance of FICA taxes would be reduced.
I recognize that political, privacy and budgetary concerns work against early creation of such a universal system. But Congress should authorize pilot projects involving the departments of Justice and Health and Human Services to further refine the concept and draw on the experience of other on-going projects for enhancing ID systems.
Congress should promptly authorize
a pilot project to work with
interested states in developing
uniform drivers licenses...
secured by a biometric identifier...
NEAR TERM IMPROVEMENTS A CALL-IN
SYSTEM AND SECURE DRIVERS LICENSES
In the interim, the following measures, more politically and administratively feasible, would help restore the credibility of IRCA and enhance the general quality of identification in the country, most of them involving use of the unique resource of the social security number
? The INS should begin without delay the pilot project for phone-in verification of worker eligibility, testing it among a variety of employers and conditions. Reliance on the A-number on the INS data bank, the Central Index, entails serious limitations and will impede the ultimate evolution of a far more efficient, universal system based on social security numbers. But as an interim measure, all individual files on the INS central index should contain the alien's social security number as well as A-number, appropriately validated by the Social Security Administration.
? Congress should promptly authorize a pilot project to work with interested states in developing a uni-form drivers license and state ID card secured by a biometric identifier, such as a digitized fingerprint or photograph, and by a social security number of the bearer pre-validated by the Social Security Administration. The project should build on the innovations for data storage in the new California drivers license, moving ultimately to the use of smart cards. The drivers license could be appropriately coded for use in employment trans-actions in proving both identity and eligibility to work in the US. A simple binary code would specify either that the bearer was unconditionally eligible to work in the US, or that the document was valid for work eligibility only when presented in conjunction with a suitable INS document. Personal data on individual drivers would remain under the control of the license-issuing states, but available for rapid exchange among the states upon legitimate request through a central clearing house.
The existence of numerous models, precedents and preparatory studies would permit rapid progress toward secured licenses and non-driver ID cards for the nation's nearly 180 million users. The Commercial Drivers Licensing System (CDLS) now serves as the model for secure licenses using biometrics. Covering five million truck and bus operators, CDLS has a central clearing house for sharing of data among the states to prevent multiple licensing and to track problem drivers.
The social security number now is used for drivers licenses by 41 states. In its 1988 report to Congress on the feasibility of a system of employer validation of social security numbers, the Social Security Administration counterproposed a system of pre-validating social security numbers on licenses for the states. With costs and political considerations impeding the development of a secure social security card, the secure drivers license or state ID card with a pre-validated social security number would become a useful surrogate in transactions where confirmation of the social security number is needed. The states would also thus be positioned to ensure the integrity of the voter rolls if Congress were to enact Motor/ Voter legislation as originally considered in the 101st Congress.
Finally, the concept of pre-validation of social security numbers should be considered for use in other vulnerable identification systems, such as passports issued by the Department of State. The machine-readable personal data page of the US passport, with pre-validated social security number and a biometric identifier, and with secure lamination and other tamper-proof features, could be produced and sold to passport holders as a secure, voluntary ID card. Such cards, uncontestable evidence of identity and work eligibility, could be an important revenue raiser and an alternative document for travel within North America and Western Europe.
Thank you.