Power is the ability to persuade others by overt or covert means to do one’s will.
There are many different ways that the computer enhances the power of organizations. With the computer, organizations can collect large amounts of information about the current and past activities of individuals and groups. With the computer, organizations can improve the way they file the information they collect and thus increase the speed and ease of reviewing it. With the computer, organizations can analyze information about the activities, opinions and social characteristics of individuals in ways that allow them to anticipate the future actions, desires and fears of the groups of people with whom these individuals associate.
The ability to make statistically reliable predictions about the future behavior of people is significant. Most computer experts, government officials and corporate pooh-bahs judge this ability benign, a tool to help them improve their communications with their customers and constituents and increase the efficiency of their organizations.
Something more malevolent, however, may be at stake. The ability may be used to allow the people in control to say only what they already have determined their listeners want to hear. The ability also may be used to develop a series of different, but not necessarily inconsistent, statements about a problem that raise complicated ethical questions about the nature of truth. The ability, in short, can be used for cynically manipulative purposes that tend to undermine the democratic process.
Because this particular application of the computer requires a large amount of sophisticated expertise and equipment, it usually is available only to the richest and most powerful institutions of our society. The isolated citizen sitting in front of his personal computer cannot be a player in this very special league.
To the presidents, corporate executives and often unknown bureaucrats who direct the great organizations of our day, the ability to anticipate the swings of mood of different segments of the population is clearly of great assistance when they seek to persuade others to buy their soup or vote for a particular candidate. It is not for nothing that Richard Wirthlin, the president of a company called Decision Making Information, did nearly $1 million worth of polling for the White House and the Republican National Committee during the first year Ronald Reagan was in the White House. According to Patrick Caddell, another pollster, this was about four times more than the Democratic National Committee paid him during the first year of Jimmy Carter’s presidency.
But the computerized techniques of power involve much more than polling. Consider, for example, the state of Missouri in 1978. No major offices were up for grabs that year, but an issue had been placed on the ballot that stirred intense feelings among many voters. The leaders on one side of this particular battle were a group of businessmen who had decided the time was ripe to challenge a fundamental source of union power.
Ordinarily, when a new worker is hired by a unionized company, he is required to join the union, at least to the extent of paying monthly dues. This is required by federal law. But under this same federal law, state legislatures are permitted to modify the union-shop requirement. Twenty states so far have done this, passing right-to-work laws that eliminate mandatory union dues. The businessmen of Missouri hoped their state would become the twenty-first to take such action because the early polls showed that about two-thirds of the voters supported the right-to-work initiative.
The union leaders knew that if the initiative was approved their bargaining power would be significantly weakened. They convened a war council. To spearhead this difficult campaign they decided they needed the expert assistance of three highly skilled political tacticians from Washington, D.C.
The experts were Matt Reese, a long-time Democratic political consultant; William Hamilton, one of the country’s more sophisticated pollsters; and Jonathan Robbin, a sociologist who over the past twenty years has developed a computer-powered marketing technique that he calls “Geodemographics.”
Together, during the summer and fall of 1978, these three men joined forces with the labor unions of Missouri and waged a campaign that illustrates the unusual political power of the computer for those who have the money and wit to hitch this powerful new tool to the particular cause of their choice.
Jonathan Robbin is a short, intense, opinionated man who does not hesitate to speak his mind. While usually polite, even likable, he occasionally erupts with a furious burst of rhetoric condemning the academics and government officials who do not fully appreciate the elegance and strength of his marketing process. Despite his outbursts, Robbin somehow suggests a professional boxer who loves to do battle but who seldom gets truly angry at his sparring partner.
“I am primarily a scientist, that is my calling,” said Jonathan Robbin as he sat in his fogbound office just across the Potomac from Washington, D.C. “By scientist, I mean I am interested in the problems of measurement and interpretation and the concept of understanding how things work and using that information and knowledge for the benefit of humanity.”
Among the organizations willing to pay for Jonathan Robbin’s kind of humanistic science are Time, Inc., General Motors, Sears, Reader’s Digest and the U.S. Army. The 1978 campaign to persuade the voters of Missouri to reject the right-to-work provision favored by big business was the first time he had applied his science to selling a complex political message.
Like almost all concepts that are effective, Geodemographics is based on a few very simple perceptions. A key insight, Robbin explains with no sign of embarrassment about the hackneyed phrase he chooses to illustrate it, is that “birds of a feather flock together.”
Translated from the cliché to the real world, Robbin says, this means that despite the penchant of the American people to move from one area to another, the economic and social characteristics of most American neighborhoods are remarkably stable.
Robbin’s second and directly related insight is that because of this stability, the information collected every ten years by the Census Bureau can be used to define the social class and economic status of people living in most of the country’s neighborhoods for the entire decade.
With these observations about the significant continuity of American neighborhoods and the nature of the decennial census, Robbin set to work. By using his computers to analyze thirty-four bits of information about the income, education and makeup of the approximately 900 anonymous people living in each of the 240,000 census units of the United States, the scientist-businessman was in a position to place each of these units on the rung of a gigantic social ladder.
But even in the computer age, comprehending a 240,000-rung social ladder is a challenge. So Robbin simplified his system in two ways. First, he reduced the number of districts by assigning the 240,000 census units to the 36,000 postal zip code areas in which they were located. Second, instead of ranking each district on a continuous scale, he shortened the ladder by devising a social hierarchy with forty specific rungs.
Robbin dreamed up catchy, easy-to-remember names for each of the rungs in his truncated ladder. He calls one rung Blue Blood Estates, another Furs and Station Wagons, a third Bohemian Mix. For the less elevated rungs, he has chosen names like Bunker’s Neighbors, Urban Renewal, Tobacco Roads and Hard Scrabble.
About 0.5 percent of America’s households, for example, live within the boundaries of the zip code areas that the sociologist has nicknamed Blue Blood Estates, the top rung of his ladder. He defines these zip codes as consisting of “America’s wealthiest social economic neighborhoods, populated by super-upper established managers, and heirs to old money, accustomed to privilege and living in luxurious surroundings.”
Bohemian Mix contains about 1 percent of the American households and includes the zip codes that cover such neighborhoods as Greenwich Village in New York, Dupont Circle in Washington, Hyde Park in Chicago and Russian Hill in San Francisco. The Bohemian Mix zip codes, Robbin explains, “have a high concentration of people in the entertainment field, and the graphic and communication arts, as well as students and educators. They are marked by well-educated young people with few children and frequently are adjacent to center city universities.”
The zip code areas included in the cluster nicknamed Bunker’s Neighbors is a mix of urban middle-class row- and apartment-house neighborhoods that contains 5 percent of all U.S. households, one of the largest groups in his taxonomy. Often bordering on urban fringe industrial parks, the residents tend to be second-echelon managers, technicians and skilled blue-collar workers.
“Humans group themselves into natural areas where the resources—physical, economic and social—are compatible with their needs,” Robbin explained. “They create or choose established neighborhoods that conform to their lifestyle of the moment.”
While neighborhoods thus tend to considerable stability in their overall makeup, the individuals living in them may be quite mobile. “As people’s lives change, they may move on to more accommodating surroundings,” the businessman-scientist said. “In their twenties, they typically marry and choose relatively transient rental housing. In their thirties, they buy housing suitable for raising small children. In their forties they are launching teenagers into society and in their fifties they are settling down to increased community participation—there are proportionately more registered voters in this age group than any other. During their sixties and beyond, they seek out domiciles suited to leisure and retirement.”
The intricately detailed base of knowledge about exactly where different kinds of people live that has been developed from census information by Robbin and his staff is then meshed with a somewhat enlarged but conventional public opinion poll. The result: commercial and political salesmen no longer have to design their campaigns on the basis of polls that only measure the opinions of men or women, rich or poor, black and white.
“This means that a whole state can be polled with a sample of about 1,000 to 1,500 people and the results extrapolated through the cluster system so that the small component areas of the state will appear to have been completely polled,” Robbin said. Because the computer has been instructed in which zip code area, ward, precinct, congressional district, township, county and city every census tract is located, accurate estimates of the proportion of those who are for or against or undecided about a given candidate or product can be presented at whatever level is desired by the customer.
“As a result, direct mail, phone or personal contact can be aimed at consumers or voters with a degree of selectivity hitherto unknown,” Robbin said.
The 1978 right-to-work battle between big business and big labor in Missouri was the first time Robbin mobilized his Geodemographic data base for directly political purposes. “The early polls in July of 1978 showed the right-to-work initiative passing by a 60 percent to 40 percent margin,” Robbin said. “After labor hired Matt Reese, Bill Hamilton and me—specifically the use of our marketing systems—the voters went to the polls and defeated the amendment by an overwhelming majority, 59 percent no to 41 percent yes.”
What had Reese, Hamilton and Robbin done to help achieve this remarkable victory? First, a computerized list of 1,467,823 telephone households in Missouri were purchased from a commercial supplier. Then, by means of a computer, each household was labeled by social class according to the location of the household in any one of the state’s 6,104 census units. In some areas the computerized lists of registered voters were matched with the list of clustered telephone households. A random sample telephone poll was conducted. The poll was large enough so that the views about the right-to-work issue of each of the state’s social clusters could be predicted.
The team working for the United Labor Committee of Missouri then divided the state’s voters into four primary groups. One group included the names and numbers of all the people living in zip code areas where the poll had shown that most voters favored labor’s position. The second group includes the zip code areas where most of the voters indicated they might listen to labor’s argument. The third group were the zip codes against labor. The final group contained the clusters where the polls showed a substantial number of people would not vote at all.
Because get-out-the-vote efforts were least likely to succeed in areas where the residents were opposed to labor or did not intend to vote, the political strategists discarded these people as a lost cause.
Again using the computer, the technicians obtained the telephone number and address of every household located in the clusters that the polls had shown contained the largest proportion of favorable or persuadable voters. The computers printed out phone-bank call sheets. The labor unions recruited the large number of volunteers necessary to operate the phone banks. The volunteers made 536,000 calls to identify precisely where each household in the target clusters stood, made an initial sale pitch and signed on block captains and new volunteers to the cause.
Based on the information collected during the original poll and the massive telephone survey, the team of Reese, Hamilton and Robbin again used the computer to print and mail eighteen different letters about why the right-to-work law should be defeated. One version of the letter, for example, was tailored to answer the questions of Missouri farmers, who the pulse finders had determined were willing to listen to labor’s message. The other seventeen letters, while not providing any directly conflicting information, were carefully designed to appeal to the other specific economic and social groups that labor was attempting to influence.
The complex campaign mounted by the Missouri labor movement and its hired strategists—Matt Reese, William Hamilton and Jonathan Robbin—was extraordinarily effective. The right-to-work initiative was defeated by a two-to-one margin on election day of 1978. The abrupt turnaround shocked a lot of politicians and reporters.
Tom Eagleton, for example, is the Democratic senator from the state and a strong opponent of the right-to-work proposal. But he was sure that labor was going to lose. Just two weeks before the election he told a reporter—Dana L. Spitzer of the St. Louis Post Dispatch—that “we’re going to get slaughtered.”
Spitzer, in an interview, said the key to labor’s victory was getting out a lot of people who were not normally labor supporters and thus not on the computerized mailing list developed from the membership rosters of the Missouri unions. A second important factor was timing. “Because labor restricted their campaigning to direct mail, telephone and door-to-door canvasing until the very end of the campaign, the business types didn’t see what was coming at all. Neither did we in the press.”
Not everyone agrees that labor’s use of Geodemographics was the critical factor in its successful campaign. Sam Landfather, who was a spokesman for the Right to Work Committee, said the principal reason the provision was defeated was the decision of several large corporations not to become involved because of threats by labor. Frank W. Emig, the assistant director of the AFL-CIO’s Department of Organization and Field Services, agrees with Landfather to the extent of playing down the importance of Geodemographics. “Labor had done an awful lot of organizing work before Matt Reese and Jonathan Robbin came to Missouri and those phone banks and neighborhood canvases could not have been done without our people,” he said.
But the volunteers would not have known whom to call and where to canvas without the sophisticated use of computers, and Jonathan Robbin has gone to work in a number of other campaigns, including the 1980 drive to reelect Jay Rockefeller the governor of West Virginia.
Although Robbin is understandably enthusiastic about Geodemographics, even he acknowledges that it poses potential problems for the democratic process. “I am concerned about using knowledge and information for the benefit of humanity. That’s not always true of scientists, those who have built weapons, bombs, all that. But it is absolutely true that an instrument for good can be turned around and used for evil. We, humanity, obviously have a sort of self-destructive tendency, and I appreciate the problem this entails.”
But Robbin was unable to suggest an institutional safeguard to protect society from the potential abuses of the high-powered process he perfected and other entrepreneurs are now copying. “I have discretion about where Geodemographics is used and I exercise this discretion according to my conscience. As the proprietor of this company, I can always say no to any client. If the Nazi party came to me and said they wanted my help in persuading the American people to exterminate the Jews, I would report them to the FBI. I don’t have to work for anyone I don’t want to.”
There are, however, more subtle problems than putting Geodemographics to work for the Nazi party. One such problem, common to all systematic targeting programs, is that they tend to encourage individual candidates or special interest groups such as the United Labor Committee of Missouri to systematically exclude voters who live in areas where the polls have determined there are a relatively large proportion of people who will not go to the polls. From the point of view of the individual candidate or group with a limited budget, of course, the process of concentrating all propaganda efforts on those who are favorably disposed or persuadable makes eminent sense.
“One of my biggest concerns is that we are playing into voter apathy,” said Michael McAdams, a tall, likable political operative in the Matt Reese organization who is widely known by the nickname Goose. “With the exception of the occasional voter registration drives, most political campaigns now work on the assumption that you will throw out the nonvoters. You always concentrate your limited budget on the undecideds and the soft-pros. You don’t bother trying to speak to the 50 percent of the population who aren’t going to vote. I am afraid this polling technique, while completely sensible from the point of view of the individual candidate, may be hurting the United States by decreasing voter participation. We no longer try to package candidates, we package audiences.”
A second problem related to targeting is that the hardware, data bases and expertise required to identify the clumps of people of greatest interest to a particular candidate or cause also can be used to communicate with these selected populations in a very special way. Consider, for example, how Jay Rockefeller used Geodemographics in his 1980 campaign to get reelected the governor of West Virginia.
“We went in with our extensive polling based on Jonathan’s Census Bureau data and determined that coal was the overwhelming interest of the people of West Virginia,” recalled Goose McAdams. “But the polls showed coal meant very different things to different people. If you were a working man, coal meant jobs. If you were the owner of a mom-and-pop store, coal meant workers with money to buy your groceries. If you were an academic type, coal could mean several things but the issue of energy independence usually was of interest to you. So everywhere that Governor Rockefeller campaigned, during the fall of 1980, he talked about coal. But when he was talking in working-class neighborhoods, he talked jobs. When he was talking in middle-class areas, he talked about improving the economy, and when he was talking at the universities, he talked about coal helping the United States achieve energy independence.”
Is there anything inherently wrong with this? In the democratic process of a small New England town, every candidate for office is known to all the voters in the village. Because the candidates and the voters know each other, work together and shop together, the candidates rarely are able to make conflicting or inconsistent promises. That is the ideal. In the larger state and national elections, reporters have come to serve as surrogate scorekeepers for the voters, score-keepers who can write a story about the campaign speech in the South that conflicted with the campaign speech in the North.
But in an age when computerized direct-mail machines are spitting out scores of different but not necessarily inconsistent messages directly aimed at the mailboxes of millions of prescreened voters, the truth about a candidate and his promises may be obscured for both the individual voters and even the most aggressive reporters.
Politicians often make private promises to groups or organizations in order to win their backing. On October 20, 1980, for example, Ronald Reagan, then a presidential candidate, wrote a letter to the Professional Air Traffic Controllers Organization promising that if elected he would “take whatever steps are necessary to provide our air traffic controllers with the most modern equipment available and to adjust staff levels and work days so that they are commensurate with achieving a maximum degree of public safety.” Neither candidate Reagan nor the controllers chose to make this important policy decision public. The controllers, however, did make public their decision to fully support Mr. Reagan’s election, and then they learned that presidents do not always keep their promises.
Republicans are not the only ones to make private promises for public support. Candidate Jimmy Carter promised the national fire fighters union that he would back a change in the labor law reducing their working hours, and for the first time in its history the union endorsed a candidate. When he got in the White House, however, President Carter reneged on his promise and vetoed the bill giving the fire fighters their requested relief.
Despite Jonathan Robbin’s admission that his technology could be abused, he contends that it does not encourage politicians to put on one mask for voting group A and another mask for voting group B. “I have found politicians remarkably honest in this sense,” he said. “None of the politicians I have worked with—and I include the special interest groups—have ever suggested or even wanted to say one thing to one person and another thing to another person in a way that would conflict with their own feelings or point of view. None of them have ever suggested taking internally inconsistent positions for the sake of getting elected.”
Dr. Walter Maner, a professor of philosophy and computer science at Old Dominion University in Norfolk, Virginia, does not share Robbin’s positive views about the new techniques of computerized politics. “Selective disclosure threatens the foundation of autonomous action,” he contends. “This is because filtered truth, though still the truth, reduces the quantity of information available, which in turn reduces our perception of the options. And since we can only make choices to the extent we are aware of the alternatives, it is easy to see how selected disclosure places limits on the exercise of choice.”
He adds that if the intent of a candidate in making different statements to different segments of society is to misrepresent or conceal the truth, then the practice is the moral equivalent of lying.
Many politicians and political buffs dismiss the notion inherent in Dr. Maner’s criticism that the direct mail targeting made possible by the computer may be inflicting serious wounds on the democratic process of the United States. “Politics is politics,” they say. “We used to have ward heelers, now we have consultants, and I don’t see any important difference.”
David S. Broder, the highly respected national reporter for the Washington Post, disagrees. “There has been considerable change since 1960, real change,” Mr. Broder said in an interview in the curious echoing glass cubicle that serves as his office in the middle of the Post’s newsroom. “The way national candidates and special interest groups are now using the tools of targeting tends to lessen accountability,” he said. The reporter also said he agreed with the thesis of Goose McAdams and others that as used so far, modern computer techniques seem to be increasing the number of nonvoters. Mr. Broder believes that this trend might possibly be reversed with reawakening of the Republican and Democratic parties, who by definition have far broader goals than the election of a single candidate.
At least in theory, any political candidate or party or cause has equal access to the targeting machinery offered by the computers of Jonathan Robbin and his competitors. But as noted in an entirely different context by George Orwell, some are more equal than others.
One particular group that is uniquely more equal than almost anyone are the members of the United States Congress. During the last few years, for example, they have voted to spend millions of dollars of public tax money to build themselves a wondrous computer system that provides them with all sorts of handy services. House and Senate members can keep abreast of the exact status of any particular bit of legislation by tapping out a few inquiries on a computer terminal located in the office of each and every one of them. They can also locate relevant studies of controversial issues that may have been prepared by the Library of Congress, the General Accounting Office or other expert fact-finding organizations. The congressional computers thus can be used to increase the knowledge and power of individual members who are seeking to keep track of complex social issues, the legislative activities of special interest lobbying groups or the program of the lumbering bureaucracies of the executive branch. When used for these purposes, the congressional computers can be seen as tending to right the balance between the David of the legislative branch and the Goliath of the federal agencies.
But the elegant, tax-supported computers of Congress are used for another purpose: helping individual members of the House or Senate stay in office. When combined with a massive subsidy that allows each member of Congress to mail a steady flow of self-serving literature at the expense of the public, the congressional computers give incumbents a significant financial advantage over their challengers. The computers and the subsidized mass mailings made possible by a privilege called “the frank” thus undermine democracy by making it hard for individual citizens to challenge the beliefs and policies of those already holding office.
A few years ago, for example, Senator Jacob Javits placed a direct mail expert named Lee MacGregor on the payroll of his Senate office in Washington. It was primary time again, and like all senators, Mr. Javits was authorized by law to harness the Senate computers and the U.S. Postal Service to prepare and deliver to his constituents just about any letter he decided was appropriate.
It was Lee MacGregor’s task to prepare a strategy by which the computer and Postal Service could be most effectively used to reelect his boss. In a preliminary memorandum to Senator Javits, MacGregor noted that among the lists of names “on our computer” were 22,000 Republican county and state officials, 40,000 federally licensed pilots, 400,000 persons with state licenses for such occupations as selling real estate and cutting hair, and an Agriculture Department list of 75,000 New York farmers.
Because of Senator Javits’ image as a Wall Street lawyer who was primarily interested in urban problems, MacGregor recognized that the computerized Agriculture Department list was important to winning the primary. What was needed, he thought, was a gimmick that would demonstrate Javits’ deep commitment to an issue of significance to the folks down on the farm.
After several months of searching, the direct mail expert found the gimmick he was looking for: a bill that had been introduced some months before by a midwestern senator to improve the safety of school buses. MacGregor persuaded Senator Javits to add his name as one of several co-sponsors to the legislation already proposed by his colleague.
In July of 1973, Mr. Javits sent a folksy two-page letter to most of the 75,000 New York State farmers on the Agriculture Department list. “I wish you could see some of the school bus accident statistics and reports I have spread out on my desk,” he began. “For me, as for many parents, these figures are more than statistics, as my own daughter was once injured in a school bus accident. They prompted me to introduce legislation directing the Secretary of Transportation to establish strict school bus safety standards,” he added, deftly inflating his role as co-sponsor to sponsor.
Senator Javits won the primary and subsequent general election, thanks in part to the strategic skills of his direct mail expert, the collating and printing power of the Senate computer and the massed muscle of the Postal Service.
Two questions are raised by the computerized campaign tactics of Senator Javits and of virtually all incumbent congressmen who seek reelection. First, should the taxpayer be required to provide his congressman with a substantial involuntary campaign contribution in the form of thousands of dollars worth of computer time and Postal Service deliveries? Second, does the ability of the computer to print 15,000 individually tailored letters in a single day encourage the members of the House and Senate to leap on issues merely because they possess a computerized mailing list and not because they have identified a genuine social problem?
One major obstacle to understanding the fundamental changes being wrought in our political lives is that the systematic targeting of voters by the AFL-CIO, most members of Congress and the special interest groups is in a narrow way as old as politics itself. It seems likely, for example, that a toga-wearing elder in ancient Greece sometime or other anticipated the noble words of Barry Goldwater and declaimed: “You only go duck hunting where the ducks are.”
But the speed, size and versatility of today’s computers have carried the up-to-date political operator a long way from the Republic of Athens. Consider, for example, Sacramento, California, in the year of 1980. Sacramento is the capital of California and at that time was the base camp of Governor Edmund G. Brown, Jr., a political mystic of some note. In December of that year, W. B. Rood and George Reasons, two excellent reporters with the Los Angeles Times, sent their paper a cracking good story. Governor Brown was using state funds to compile a massive list of political backers that was to be cranked into a sophisticated and highly computerized mailing system that also was being paid for by the state.
In their first story, the two reporters wrote that a cadre of seven persons already had gone through 750,000 letters sent to Governor Brown and compiled cards containing the names of and personal data and background information about 50,000 political activists. The name cards had been prepared for use in an advanced computer system that California had leased from the Xerox Corporation.
“Our new record processing system has 130 [storage category] fields,” the Los Angeles Times quoted one official in the governor’s office as saying. “The first three or four are for names and addresses, then all the rest are for things like antinuclear, environmentalist, or whatever. Then you can tell the machine, ‘I want to write all the women that are for abortion or against abortion.’ The machine will go through and pick out all the women that fit that description, even those just in a certain zip code. The machines also can store form letters for specific purposes such as answering people opposed to or supporting a particular piece of legislation.”
Two weeks after the printing of the story, at a news conference in Los Angeles, Governor Brown said the new system had not been developed for political purposes but that to insure its integrity the names of political supporters would be removed from the computer. “This was an attempt to create a two-way communication instead of an isolated entity that I thought my office had become,” he told reporters. “It was done with the best of motives.”
But back in Sacramento, Mr. Rood and Mr. Reasons discovered some rather convincing evidence suggesting that Brown’s motives were not quite as he described them. The new evidence was a confidential memo showing that the seven-person task force and the computer system were part of a master plan that called for using the governor’s “entire staff” to put together a national political organization for Mr. Brown.
The memorandum, which was written on November 14 in the state’s office in Washington, D.C., called for the mobilization of the governor’s staff “to develop a national constituency to assist in refining the governor’s themes and thereby expand the base of his support.”
This goal was to be achieved by putting together dossiers on people identified as “suspects,” a term used in the memorandum to refer to people whose backgrounds and views were under study to determine their usefulness as political contributors and volunteers for Governor Brown. “To tap the larger pools of the universe it is essential that every member of the staff participate to the fullest extent in culling names from various sources.”
The staff, the memorandum said, would be given cards to fill in the names of “suspects” and “will be responsible for doing preliminary research which will include tracking down the addresses, occupations, employers/affiliations, sources and contacts on those people we intend to put in our outreach program.”
When asked about the memorandum, the governor’s chief of staff told the Los Angeles Times reporters that six employees had been fired after Brown learned of the master plan and that steps had been taken to make sure that “neither the letter nor the spirit of the memo would be implemented in this office.” The sincerity of this second promise to avoid using a state computer to build a national constituency for Governor Brown was put into question a year and a half later in a report by the California Fair Political Practices Commission.
Top campaign aides to Brown, the report said, had sought to block the commission’s investigation by destroying, concealing and altering important evidence. While Brown himself was not accused of misconduct, the report said he too had failed to make evidence available as he had promised when he testified under oath about the matter.
Politicians like Jerry Brown use computers to enlarge their personal power. Corporations, government agencies and other bureaucracies use computers to enlarge their institutional powers.
The Internal Revenue Service is an example of one of the latter. Before turning to how computers have been harnessed by the IRS during the last few decades, consider its legal mandate. In the knowledge that collecting taxes is vital to the government, the United States Congress and the courts gradually have granted the IRS an extraordinary legal authority which in many ways exceeds that handed to any other federal enforcement agency. The power begins, of course, with laws that make it a serious crime for almost every household, corporation and other kind of institution not to file an annual report describing its yearly income, the sources of the income and, to a somewhat lesser extent, where the income was spent. For the individual or institution whose income is derived from a single source, this annual report can be quite brief. But for those receiving dividend checks, interest, commissions, rents and other forms of income, the reporting requirements are incredibly broad.
The power of the IRS is enhanced by the authority of the agency to make very broad inquiry of the taxpayer to verify tax returns. Under the law and the interpretations of the Supreme Court, the IRS has the unique power to issue a summons to a taxpayer, employer, accountant or any other third party. The authority of the IRS to issue a summons on its own authority contrasts with other law-enforcement agencies such as the FBI, who must apply to a court when they wish to compel the presence of a person of interest.
The precise reach of an IRS summons is subject to some controversy. According to a report done for an independent federal study group a few years ago, a citizen must appear when served by an IRS summons, but does not actually have to produce the requested papers until compelled to by a judge. But as noted in the report prepared by a team of legal scholars headed by Charles Davenport, then at the Law School of the University of California at Davis, the IRS does not tell the taxpayer that only a judge can order him to produce his records. Instead, the report said, the IRS gives the taxpayer an “uninformative and misleading” statement that “wrongfully implies that the summoned party must appear and give testimony or produce documents or penalties will follow.”
Over the years, the detailed personal information collected about millions of Americans by the IRS has made the agency’s files an almost irresistible temptation to those who could gain access to it. The tape-recording system installed in the White House by President Nixon gave the world a unique indication of the pull of this particular magnet.
“Do you need any IRS (unintelligible) stuff?” Mr. Nixon asked his assistant, John Dean.
“… We have a couple of sources over there I can go to,” Dean replied. “I don’t have to fool around with Johnnie Walters or anybody, we can get right in and get what we need.”
This particular exchange has a unique place in American history because it was part of the evidence supporting one of the proposed articles of impeachment against Mr. Nixon. The president, the article said, “acting personally and through his subordinates and agents, endeavored to obtain from the Internal Revenue Service, in violation of the constitutional rights of citizens, confidential information contained in income tax returns for purposes not authorized by law.”
The subsequent Senate testimony of John J. Caulfield and Clark Mollenhoff, both White House assistants, and Roger V. Barth, an aide to the head of the IRS, indicates there were at least fourteen occasions when information was drawn from the IRS files at the request of someone in the Nixon White House. Individual subjects included George Wallace, his brother Gerald, the Reverend Billy Graham, the actor John Wayne and a candidate for a post in the campaign to reelect Mr. Nixon. While there are no records of such White House requests during either the Eisenhower or Johnson years, there is a record showing that Carmine S. Bellino, a consultant to President Kennedy, examined an unknown number of returns and associated documents in the spring of 1961.
But the abuse of the information that every American must file with the IRS under the penalty of law was not limited to the White House. In 1965, for example, FBI agents obtained the tax returns of several Ku Klux Klan leaders. The tax information, Senate investigators later learned, was used by the agents in what the FBI called COINTEL, a formal and secret counterintelligence effort to “disrupt” groups and “neutralize” individuals who had been judged a threat by bureau director J. Edgar Hoover. In this case, according to an FBI memorandum of May 10, 1965, the purpose was to “expose” the selected Klan leaders to the members of their organization and the public “by showing income beyond their means.”
The investigative power of the IRS has focused on far less sinister targets than the Ku Klux Klan. In early 1967, for example, the Central Intelligence Agency heard rumors that a magazine called Ramparts was about to publish an article about the CIA’s direct involvement in the funding of the National Student Association. On February 2, 1967, a CIA official met with the assistant commissioner of the IRS and several of his aides. “I suggested that the corporate tax returns of Ramparts be examined, and that any leads to possible financial supporters be followed up by an examination of their individual tax returns,” the CIA man wrote in his action report.
The CIA request and the IRS response were an astounding example of how two established government agencies attempted to frighten a magazine in a way that appears to fly in the face of the First Amendment of the Constitution. Until 1969, however, the use of the IRS tax data for political purposes seemed somewhat limited and even erratic. That year, however, the IRS established the Special Service Staff (SSS) to systematically gather intelligence on a category of taxpayers mostly defined by their political activities.
According to a much later report by the Senate Select Committee on Intelligence Activities, the SSS was formed as a result of pressure from a group of senators concerned about organized crime and subversion and from President Nixon, acting in part through Dr. Arthur Burns, then the chairman of the Federal Reserve Board. According to a memorandum written on June 16, 1969, Dr. Burns informed the head of the IRS about the president’s concern “over the fact that tax-exempt funds may be supporting activist groups engaged in stimulating riots both on the campus and within our inner cities.”
The Senate report said its investigators had failed to uncover any evidence suggesting that either President Nixon or Dr. Burns had provided the IRS with evidence to support the assertion that such tax-exempt groups were violating tax laws.
Immediately after the IRS created the SSS, the Justice Department and the FBI provided it with shopping lists of 2,300 organizations categorized as “old left,” “new left,” and “right wing” and a computer printout of about 10,000 individuals who had been identified as being active in the civil rights and antiwar movements. The computer printout of the activists had been compiled a year or so earlier at the request of Ramsey Clark, the Johnson administration attorney general who later became an outspoken opponent to the Vietnam War and a strong advocate of many liberal causes.
“The SSS opened files on all these taxpayers, many of whom were later subjected to tax audits and some to tax fraud investigations,” the committee found. There was no evidence, it continued, that the names on the Justice Department and FBI lists subjected to IRS investigations “were selected on the basis of probable noncompliance with the tax laws. Rather, these groups and individuals were targeted because of their political and ideological beliefs and activities.”
Names on the SSS list included Nobel Prize winner Linus Pauling, senators Charles Goodell and Ernest Gruening, Representative Charles Diggs, journalists Joseph Alsop and Jimmy Breslin, and Washington attorney Mitchell Rogovin. Organizations included political groups ranging from the John Birch Society to Common Cause, professional organizations such as the Legal Aid Society, publications including Playboy and Commonweal, and government institutions such as the United States Civil Rights Commission. No explanation has ever been provided why an agency that naturally pays no taxes would be of interest to the IRS.
By the time the SSS was abolished with the first breath of publicity in 1973, the unit had gone over about half of the 10,000 names on the Justice Department list, some of which were referred to IRS field units for formal investigation.
The IRS was not to be left naked before its enemies. During the same year that the agency was dismantling the SSS, it was establishing a central computer index—the Intelligence Gathering Retrieval System—for collecting general intelligence data, a good deal of which is not directly related to tax law enforcement. Within a few months, more than 465,000 Americans were indexed in the new system, including J. Edgar Hoover, the IRS commissioner and thousands of others not suspected of any tax violation. Under the system, intelligence gathering was begun on an individual before the IRS had received any specific allegation of wrongdoing. The sole criteria for deciding whether to add a name to the computerized file was possible “future value.”
But the computers of an agency as large and powerful as the IRS do not have to be specifically dedicated to intelligence gathering before they become a matter of concern. In 1976, for example, the IRS proposed to Congress that it approve the Tax Administration System (TAS), a new computerized network that the House Appropriations Committee estimated might cost $1 billion.
Because of the great size and complexity of the TAS network, Representative Al Ullman, then chairman of the House Ways and Means Committee, and Representative Charles Vanik, chairman of the Oversight Subcommittee of the Ways and Means Committee, worried that the new IRS computer might grow into a “system of harassment, surveillance and political manipulation.” The two men joined in requesting that the Office of Technology Assessment, an arm of Congress, undertake an independent analysis of the network.
After a lengthy investigation, the OTA concluded in its windy but thoughtful way that TAS would have an enormous impact. The system, the report said, would “determine or affect the collection, use, maintenance and dissemination of large amounts of information about citizens. It will play a pivotal role in governmental and private data banks and information systems which contain the details of the personal, organizational and business lives of Americans at home or abroad.”
The OTA made a long list of the attributes of TAS that it thought Congress should consider. The system, the investigators said, would decentralize taxpayer files and make them instantly available to those who share and use federal tax information. It would create a national system for quickly transferring these histories anywhere in the United States. It would give the IRS an expandable data base with the capacity to acquire and store more detailed histories of individual taxpayers for longer periods of time. The system, by increasing the capacity of the IRS to associate the bits of information it collected, would enable the agency to create what in a sense can be considered new information.
Because of these various attributes, the OTA concluded, the proposed computer system “could be perceived as posing a threat to civil liberties, privacy and the due process rights of taxpayers. These effects might include a potential for surveillance, harassment or political manipulation of files, for which specific controls and safeguards are of concern to Congress.”
Much to the annoyance of the IRS and many of the computer experts in government and industry, the Carter administration in January of 1978 decided too many questions had been raised about TAS and the project was scuttled. Instead, virtually all of the money the planners had estimated it would cost to build TAS was diverted to improving the IRS’s existing computer system. This time the fresh memory of the Watergate years and how the Nixon administration had sought to use computerized dossiers to harass its ideological enemies had halted a system sold under the usually invincible banner of increasing government efficiency.
Meanwhile, however, the IRS has continued to develop and refine a very different kind of computerized information system called the Taxpayer Compliance Measurement Program (TCMP). Although it is largely accepted by critics both in and out of Congress on the grounds that it improves the ability of the agency to collect taxes, Paul M. Strassels and Robert Wool in their book All You Need to Know About the IRS said the TCMP was “the closest thing to 1984 market research that our country has.”
Simply stated, the TCMP is a computer-assisted process to help the IRS predict the behavior of every American taxpayer. To the degree that it is successful, the TCMP enhances the power of the IRS in relation to the citizenry by allowing it to concentrate its small army of investigators and accountants on the targets most likely to have tried to cheat the government.
The program works like this: since 1962, a random sample of different kinds of taxpayers have been selected on a periodic basis. These are not individuals, corporations or other organizations who are suspected of any wrongdoing, just a cross section of taxpayers. Once identified, the individuals and institutions are subjected to an extremely intensive examination.
“Experienced revenue agents and tax auditors from the IRS Examination Division conduct in-depth audits of each of the sampled returns,” reported Susan B. Long in a Justice Department–financed study on the government’s response to tax evasion.
“Detailed check sheets are made out by the IRS examining officer on the amounts reported line by line on the return and the ‘corrected’ amounts after the audit,” Mrs. Long continued. “Supplemental information concerning the taxpayer’s financial affairs, who prepared the tax return and what procedures were used in carrying out the TCMP are also included.”
In one recent survey, 200 separate items were collected about each of the 50,000 individuals who had been randomly selected to represent the entire population of 93 million American taxpayers. Once collected by the auditors, the 10 million bits of information collected by the survey were fed into a giant IRS computer for analysis. The result: a line-by-line, income-level-by-income-level, region-by-region list of probabilities that a taxpayer in any one of these categories incorrectly stated the amount of tax due the government.
In addition to providing an estimate of what kinds of taxpayers might give the IRS trouble concerning what aspects of their income, the survey also provides the government a statistical indication of how much would be recovered if any particular hole was closed. Who are the real cheaters? How do they cheat? How much might the government expect to recover if it concentrated on Line 33?
From the elaborate statistical tables developed from the periodic audits of selected taxpayers, the IRS develops its enforcement strategy for the entire nation, the marching orders for its 87,000 employees. But the TCMP is more than that: it is the basis for virtually all of the agency’s basic policy decisions. The IRS uses the TCMP data to support its budget requests to Congress, to make proposals for changes in the tax law and to make decisions on such matters as staffing levels, training requirements, taxpayer education programs and the form of the tax returns.
Mrs. Long, who has a Ph.D. in sociology and is an assistant professor at the Department of Quantitative Methods at Syracuse University, has conducted research on the administration of the tax laws for many years. Because of this professional interest, she asked for the underlying data of the TCMP more than five years ago.
The Internal Revenue Service rejected her request, arguing that releasing the data would enable taxpayers to anticipate the service’s enforcement strategies.
Because of the Freedom of Information Act, however, the IRS had a serious legal problem. This law establishes the principle that any document or report prepared at public expense by a government agency must be made public unless it concerns an important military secret, the trade secret of a private company or the personnel matters of an individual citizen. The data requested by Mrs. Long clearly did not fall into any of the accepted categories, and a Federal Court of Appeals upheld her right to receive it.
In arguing the matter before the courts, Mrs. Long and her lawyer, Stephen K. Strong of Seattle, contended that a fundamental principle of American government was at stake in the refusal of this powerful agency to publish the computerized data created by the publicly funded survey. This principle is that the government of the United States is supposed to operate with the consent of the governed. Analyses of TCMP data by congressional and independent researchers, they said, “could lead to the uncovering of problems in the tax laws, tax returns and instruction forms, and administrative practices and could provide suggestions as to how to deal with these problems.”
Mrs. Long noted that the IRS often refers to TCMP data in making requests for appropriations to hire more IRS auditors. “Since the underlying data are not revealed, the public is forced to speculate on the validity of the service’s conclusions, which may or may not be valid. Analysis of the data by outside researchers could enable the public and its representatives in Congress to better evaluate the IRS’s strategy and the appropriation requests that are made to carry out its strategy.”
Despite winning the battle against the IRS at the court of appeals, Mrs. Long and her lawyer lost the war when the IRS managed to persuade Congress to pass an amendment to the Freedom of Information Act exempting what the government claimed was sensitive data from the requirement that it be made public. By preventing a knowledgeable and independent expert from examining computer-generated information that would have enabled her to judge the effectiveness of the IRS in administering the nation’s tax laws, one of the government’s most powerful agencies became even more powerful.
Thus has the computer strengthened the hand of the IRS in relation to citizens who pay taxes and citizens who seek to monitor the performance of the tax collector. The computer, however, serves to enhance the power of the management of the IRS in a third way: to help the bosses keep track of the employees.
The computer program developed by the IRS managers for this purpose is called the Audit Information Management System (AIMS). The purpose of AIMS, in the indirect language of government, is to provide the IRS with an inventory control system, timely data to make sure that its 87,000 employees are doing their job. But in the impolite language of the workplace, AIMS is a de facto quota system that requires every agent in the IRS to meet certain standards, what is known as the “plan,” to audit a sufficient number of tax returns, to identify a sufficient number of clinkers, to recommend bringing a sufficient number of criminal cases.
When General Motors imposes a quota requiring the workers in a factory to produce a certain number of cars each day, most Americans would agree that the GM managers are within their rights. But when a massive, highly computerized law-enforcement agency like the IRS imposes quotas, there can be trouble. What happens, for example, when an IRS agent knows his next promotion is dependent upon making ten cases and the luck of the draw brings him only seven? Might he be tempted to bring charges of some sort against three individuals who had not actually violated the law?
Top IRS officials have repeatedly denied that the IRS operates by quota systems. The IRS handbook for audit group managers stresses over and over again that statistics are an aid to good management, not a replacement for it. Nonetheless, an extensive investigation of the IRS for a small government watchdog agency called the Administrative Conference of the United States found that substantial numbers of IRS employees believe quotas are set and used to measure their performance.
“Group managers are well aware of the ‘plan’ and their group’s part in accomplishing it,” the report to the Administrative Conference said. “One manager, in fact, stated that he was well aware of management’s expectations for his group and he saw his primary function as assisting his group in ‘making the plan.’ This view was shared, but not so bluntly stated, by most if not all managers. Numerous interviews with agents and management personnel indicate they believe that closing cases so as to ‘make the audit plan’ is the most important part of the job.”
The first step in developing a quota system is to create a reliable mechanism to count the activities of whoever it is you want to control. For many years, for example, clerks in the offices of each of the 94 U.S. attorneys laboriously compiled lists of the investigations, prosecutions, convictions, hours in court, collections and many other activities that occupy the time of federal prosecutors. The monthly statistics were dispatched to the Justice Department in Washington, where a dreary report would be published a long time after it could assist the attorney general in directing his small army of federal prosecutors.
But the good government lobby of the federal establishment—the Office of Management and Budget, the General Accounting Office and sometimes even a Congressional committee—began to notice that the statistics did not always add up and regularly were printed many months after the events they described. Gradually the Justice Department managers decided it would help them better justify their appropriations requests if they had a way of collecting accurate and up-to-date information about the actions of the 94 U.S. attorneys and the 3,500 assistant attorneys, clerks and secretaries who work for them.
The experts decided on two converging approaches to meet the demand for better information. First, a great deal of effort was spent on deciding what data would be collected and how it would be recorded. Second, after experimenting with several approaches, a decentralized system was adopted in which the personnel in each office would file case information via terminals hooked into a small local computer.
Thomas Hagerty is a cheerful man with a broad Irish face and a small office on the sixth floor of the Federal Building in Newark, New Jersey. Three Miss Piggy posters decorate his wall and a computer terminal sits next to his gray steel desk. Thomas Hagerty is not an attorney or an FBI agent, someone you might normally find in a U.S. attorney’s office. He is instead a systems manager, the man responsible for helping a cranky old legal office adapt itself to the computer age.
“Because the information we used to send to Washington was processed on a batch basis, every month or so, and often had coding errors, it just wasn’t much help for managing the government’s business here in Newark,” Mr. Hagerty explained.
“Now, with our local on-line system, there is a tremendous immediacy of data. You can run reports by division or by attorney, you can determine how many cases so-and-so has, you can see if one attorney is overloaded or if another is not pulling his or her load. You can call up cases by opposing counsel, or defendant’s name. You can ask the computer to give you a report on cases by statute of limitations or ask it to warn a particular attorney when a lab report should be back. It is a wonderful management tool.”
The computer sits in a small locked room next to Mr. Hagerty’s office. As such matters are measured these days, it has a rather small memory, one million bytes. With seventeen terminals in Newark and one each in satellite offices in Camden and Trenton, the computer serves the needs of the fifty-five attorneys and sixty-five clerical workers. The equipment was installed in July of 1981. The Justice Department hopes that similar equipment will be installed in twenty-five additional offices within the next two or three years.
The computer systems being installed in the separate U.S. attorneys’ offices almost certainly will improve the management efficiency of the federal prosecutors. But because the tapes from these computers will be sent to the attorney general in Washington, something more subtle but perhaps more important is also happening. Without any discussion in Congress, without any change in law, the power of the attorney general to influence public policy is being enormously enhanced.
From the beginning of our government, U.S. attorneys have been somewhat independent from the attorney general. To this day, for example, no one is nominated to be a U.S. attorney without the approval of the senators in the state where the federal district is located even if the senators happen to be from a different party than the president.
The fact that U.S. attorneys are beholden to senators, as well as to the president and his attorney general, has meant that federal prosecutors could be a little bit independent from the dictates of the Justice Department. Federal law enforcement in Wyoming has not been a carbon copy of federal law enforcement in Massachusetts. Law enforcement in Oregon has had a different character than that in Florida.
Because this is a huge country, with widely varying climates, economic conditions, ethnic groups and religious backgrounds, there are some advantages in this modest flexibility. There also are some disadvantages. U.S. attorneys in some parts of the West, for example, may not have enforced environmental laws with the vigor mandated as national policy by the Congress and attorney general. U.S. attorneys in the South may not have brought an honest enthusiasm to enforcing the civil rights laws.
But with the installation of the computers in every office, the attorney general will be able to keep track of every investigation, every prosecution, every conviction and every acquittal. He will be able to do this on an individual basis—the name of a defendant, the name of a prosecutor, the name of a defense lawyer—or by broad categories of law-enforcement actions. How does the pattern of civil rights enforcement of the U.S. attorney in Louisiana compare with that of the U.S. attorney in Minnesota? Is the U.S. attorney in Missouri bringing as many official corruption cases as his counterpart in Illinois? Is the federal prosecutor in northern California paying the same attention to draft dodgers as the prosecutor in Houston?
Benjamin R. Civiletti, the attorney general in the last years of the Carter administration, believes the new computer system will be beneficial. “It was obvious that the system was going to increase the power of the attorney general—that was one of its main objectives,” he said in an interview in his Washington office. “In part, the purpose of the system is to give life to what is really the legal responsibility of the attorney general which he cannot effectively exercise now because of the semi-autonomous nature of the U.S. attorneys and the lack of information to determine whether they are following the policies of the attorney general.”
The former chief prosecutor of the United States, who now is in private practice, has an office that looks down Pennsylvania Avenue toward the Justice Department. “If the attorney general determines and publicly states that the investigative and prosecutive forces of the Justice Department are to be concentrated on certain goals, he wants to be able to determine these goals are being pursued. Let’s say the priorities are white-collar crime, organized crime, drug trafficking and espionage. If a U.S. attorney in Chicago is concentrating on stolen-car cases, then he obviously is not following policy.”
Mr. Civiletti, acknowledging that the computer system is strengthening the hand of an already powerful federal official, does not see any hazards. “Even with the computers,” he said, “there is little practical risk of an attorney general forcing the wrongful disposition of a particular case.”
Another experienced prosecutor, however, is less sanguine. Robert Morgenthau is a slim, elegant man with sharp blue eyes and carefully brushed white hair. He now is the district attorney of Manhattan. A few years ago, he was the U.S. attorney for the southern district of New York.
“The system, of course, will give the attorney general a better idea what his agents are doing and how well they are doing it,” Mr. Morgenthau said. He was sitting behind a huge conference table in his large sunny office in lower Manhattan. “But if you are going to take away or reduce the discretion of a U.S. attorney, if you are going to make him a simple instrument of the attorney general, then you are going to give an awful lot of power to the attorney general to do both good and bad. If an attorney general wanted to abuse this power, the system would give him the wherewithal. He could suppress cases, he could give out confidential information about the status of cases, he could make the Justice Department a much more formidable national force.”
The complex computer systems of the IRS nurture the growth of quotas that reduce the exercise of independent judgment of the individual agent and increase the power of Washington. The gradual development of a computerized reporting system within the Justice Department lessens the discretion of the U.S. attorneys and strengthens the hand of the attorney general.
Some contend the increased administrative controls are good, that they contribute to an evenhanded application of the law and are a way the attorney general may be held accountable. Others believe the systems are bad, that they lead to an inflexible application of the law and give Washington too much power over too many people. Jeffrey A. Meldman is a professor at the Sloan School of Management of the Massachusetts Institute of Technology. A short man with a neatly trimmed mustache and glasses, Professor Meldman has given a good deal of thought to how the new information technology is changing both business and government.
“In engineering,” he explained, “there is a principle which holds that it frequently is best to have a loosely coupled system. The problem with tightly coupled systems is that should a bad vibration start at one end of the machine, it will radiate and may cause difficulties in all parts of the system. Loose coupling frequently is essential to keep a large structure from falling down. I think this principle of mechanical engineering may be relevant to the way we use the computer in the United States.”
Professor Meldman’s analogy is an interesting one. As long as the United States is served by well-meaning, intelligent attorney generals who share the widespread American concerns about the importance of the independent autonomy of every citizen, then a tightly coupled system may best serve these goals. But should a president appoint an attorney general who had a fundamental abhorrence to organized crime figures or civil rights demonstrators or war protesters, a tightly coupled system might create vibrations that could severely damage the Justice Department, and possibly even the entire government.