The threads leading to Americans are many. The United State Congress' Privacy Protection Commission, chaired by David F. Linowes, estimated that there are over 8,000 different record systems in the files of the federal government that contain individually identifiable data on citizens. Each citizen, on average, has 17 files in federal agencies and administrations. Using these files, for example, Social Security data has been matched with Selective Service data to reveal draft resisters. IRS data has been matched with other administrative records to tease out possible tax evaders. Federal employment records have been matched with delinquent student loan records to identify some 46, 860 federal and military employees and retirees whose pay checks might be garnished. In Massachusetts welfare officials sent tapes bearing welfare recipients Social Security numbers to some 117 banks to find out whether the recipients had bank accounts in excess of the allowable amount. During the first pass some 1600 potential violators were discovered.
Computer matching and the integration of data files into a central databank have enormous ethical implications. On the one hand, the new information can be used to uncover criminals and to identify service requirements for the needy. On the other hand, it provides powerful political knowledge for those few who have access to it and control over it. It is ripe for privacy invasion and other abuses. For this reason many politicians have spoken out against centralized governmental databanks. As early as 1966 Representative Frank Horton of New York described the threat as follows: "The argument is made that a central data bank would use only the type of information that now exists and since no new principle is involved, existing types of safe-guards will be adequate. This is fallacious. Good computer men know that one of the most practical of our present safeguards of privacy is the fragmented nature of present information. It is scattered in little bits and pieces across the geography and