September 3, 2008
On election night, when the three major television networks announce the next president, the winner they announce is not chosen by the voters of the United States. He is the selection of the three networks themselves, through a company they own jointly with Associated Press and United Press International.
That company is called News Election Service (NES). Its address is 212 Cortland Street, New York City. Its phone number is (212) 693-6001. News Election Service provides “unofficial” vote tallies to its five owners in all presidential, congressional, and gubernatorial elections. NES is the only source Americans have to find out how they, as a people, voted. County and city election supervisors don’t come out with the official totals until weeks later. Those results are rarely reported in the national media.
The US government does not tabulate a single vote. The government has granted NES a legal monopoly, exempt from antitrust laws, to count the votes privately.
Those are the facts.
NES. The company is a conspiracy theorist’s dream—or nightmare.
As mentioned above, NES operates exactly the way most imaginative conspiracy theorists believe all media operate. The ABC, NBC, and CBS networks, together with the AP and UPI, own the company jointly. Associated Press is a nonprofit co-op of a large number of daily newspapers, and UPI serves many of the rest. Local television and radio stations take most of their election returns from network tabulations. NES is a very real “cabal.” Every media outlet in the United States acts in concert, at least on election nights.
NES has a full-time staff of fourteen. On election nights, that number swells to approximately 90,000 employees, most of them posted at local precincts phoning in vote totals as they’re announced. Others answer the phones and enter these totals into the NES computer. The government has no such computer. Only the privately-held NES counts the votes. I called NES’s executive direct or, Robert Flaherty, and asked him whether his company was run for profit. He wouldn’t answer. His only response was, “I don’t think that’s part of your story. ”
The company was conceived in 1964, in part as a cost-saving measure by the three major television networks (it was originally called Network Election Service), but largely to solidify the public’s confidence in network vote tallies and projections by insuring uniformity. In the California Republican primary that year, television networks projected Barry Goldwater the winner on election night, while newspapers reported Nelson Rockefeller victorious in their morning editions. The networks themselves could vary widely in their return reports.
“Many television executives believe the public has been both confused and skeptical over seeing different sets of running totals on the networks’ screens,” the New York Times reported. The networks (the two print syndicates were soon added to the setup) wanted the figures transmitted over their airwaves to be irrefutable. With all the networks—and later the print media—deriving their information from a central computer bank, with no alternative source, how could they be anything but?
“The master tally boards…would probably come to be accepted as the final authority on the outcome of races,” the Times declared. The “news media pool” was first tried in the 1964 general election. Most of the 130,000 vote counters were volunteers from civic groups. Twenty-thousand newspaper reporters acted as coordinators. NES central was located at New York’s Edson Hotel. Vote-tallying substations were set up in such select sites as an insurance company headquarters and a Masonic temple. When polls closed, the newly-formed system shaved almost 90 minutes off the time needed to count votes in the 1960 election.
News Election Service had its goal circa 1964 to report final results within a half-hour of final poll-closing time. Now, of course, they go much faster than that. In the 1988 election, CBS was first out of the gate, making its projection at 9:17 Eastern time, with polls still open in eleven states. ABC followed just three minutes later.
All of these light-speed results are, naturally, “unofficial.” County clerks take a month or more to verify their counts and issue an official tally. Plenty of time for any necessary fudging and finagling. And there may be none needed. Discrepancies are a matter of course throughout the nation’s thousands of voting precincts. The major networks rarely bother to report on such mundane matters. So who’s going to know?
One rationale behind maintaining a vote-counting monopoly is to insure “accuracy,” but in 1968, when Richard Nixon defeated Hubert Humphrey by a margin that could be measured in angstroms, the role of NES became a good deal more shadowy.
At one point in the tally, the NES computer began spewing out totals that were at the time described as “erroneous.” They included comedian/candidate Dick Gregory receiving one million votes when, the New York Times said, “His total was actually 18,000.” The mistakes were described as something that “can happen to anyone.”
NES turned off its “erroneous” computer and switched on a backup system, which ran much slower. After much waiting, the new machine put Nixon ahead by roughly 40,000 votes, with just 6 percent of the votes left to be counted. Suddenly, independent news reporters found over 53,000 Humphrey votes cast by a Democratic splinter party in Alabama. When the votes were added to Humphrey’s total, they put him in the lead. Undaunted, the Associated Press conducted its own state-by-state survey of “the best available sources of election data” (presumably, NES also makes use of the “best available sources”) and found Nixon winning again. And that’s how it turned out.
What exactly was going on inside the “master computer” at NES? The company’s director blamed software, even though the machine had run a twelve-hour test flawlessly just the day before using the same programming. Could the software have been altered? Substituted? Or was the fiasco caused by a routine “bug,” which just happened to appear at the most inconvenient possible time? At this point, its more a question of what we can know than what we do know.
With all the snafus and screwups, the real winner of the 1968 presidential election will never be certain. We do know this: Liberal warhorse Humphrey died without fulfilling his dream of becoming president, while Nixon hung around long enough to see his loyal crony George Bush in the White House. (The old, unindicted co-conspirator passed on in 1994.)
Not only does NES keep the election night vote count, but most voters cast their ballots “virtually.” Computers tabulate 54 percent of the votes cast in the United States. Sure, paper ballot elections were stolen all the time, and lever voting machines are invitations to chicanery. But there’s something sinister about computers. Though most professionals in the field, as one would expect, insist that computers are far less vulnerable to manipulation than old ways of voting, the invisibility of their functions and the esoteric language they speak makes that assertion impossible to accept.
Even executives of computer-election companies will admit that their systems are “vulnerable,” although they’re reticent to make public statements to that effect. One executive told me, right after asserting that there’s never been a proven case of computer election fraud, “there’s probably been some we don’t know about.”
Even if “we” do find out, there’s still little chance that the fraud will be prosecuted. A former chief assistant attorney general in California, Steve White, points out that without a conspirator willing to inform on his comrades or an upset so stunning as to immediately arouse suspicion, there’s little hope of ferreting out a vote fraud operation.
There are very few elections that qualify as major upsets anymore. Pre-election polling tempers the climate of opinion effectively enough to take care of that. As for turncoat conspirators, if the conspiracy works there are no turncoats. A good conspiracy is an unprovable conspiracy. It remains a conspiracy “theory.” To even talk about it is “paranoid.”
“If you did it right, no one would ever know,” said White. “You just change a few votes in a few precincts in a few states and no one would ever know.”
There are several makes of computerized voting machines. One widely used model is the Shouptronic, whose most advantageous feature is the speed with which it tabulates votes. Multiple machines can send results to a central computer instantly over land lines or by satellite.
Shouptronic is essentially an automatic teller machine for voters. All votes are recorded by button pressing. The Shouptronic leaves no physical record of votes. Like all computer vote counters, its programming is top secret.
As solid a source as Robert J. Naegle, author of the federal government’s national standards for computerized vote counting, is alarmed by the secrecy masking computer election software.
“They act like it was something handed down on stone tablets,” he says. “It should be in the public domain.”
The Shouptronic is named for its company’s owner, Ransom Shoup II. In 1979, Mr. Shoup was convicted of conspiracy and obstruction of justice relating to a Philadelphia election under investigation by the FBI. That election was tabulated by old-fashioned lever machines, which also leave no “paper trail” of marked ballots. Shoup was hit with a $10,000 fine and sentenced to three years in prison, suspended.
Another computer voting company, Votomatic, maker of Computer Election Services (now known as Business Records Corporation Election Services), emerged unscathed from a Justice Department antitrust investigation in 1981. The president of the company quipped, “We had to get Ronald Reagan elected to get this thing killed.” The remark was supposed to be a joke. Forty percent of American voters vote on CES systems.
CES machines have been described as relying on “a heap of spaghetti code that is so messy and so complex that it might easily contain hidden mechanisms for being quietly reprogrammed ‘on the fly.’” A computer consultant hired by the plaintiffs in a suit against CES described the way a CES computer runs its program as “a shell game.”
Votomatic has one especially troubling drawback. On election night 1982 in Miami, Ken and Jim Collier, who spent much of their lives tracking what they describe as a national conspiracy to rig all major elections, captured the problem on videotape. This “Votescam Video” has been the Colliers’ Exhibit A ever since. They’ve showed it to reporters at major television networks, and evangelical talk show host Pat Robertson paid them $2,500 for broadcast rights to the tape. Robertson aired a portion of the tape.
The problem with Votomatic, captured on the Colliers’ tape, is something called “hanging chad.” The perforated squares on Votomatic computer ballot cards are, for some reason, called “chad.” When a voter fails to punch it out completely, it hangs on the card.
To solve this problem and allow the computer to read the cards, election workers routinely remove hanging chad. The registrar of voters in Santa Clara County, California, says that “five percent or less” of all Votomatic cards have hanging chad, and election workers don’t pull it off unless it is hanging by one or two corners.
The vision of local ladies from the League of Women Voters deciding how voters have voted, putting holes in perforated ballots with tweezers was an astounding proposition to the Colliers. When they talked their way into the Miami counting room on November 2, 1982, toting video camera with tape rolling, that’s exactly what they found. Prima facie evidence of tampering, they believed, and Jim started shouting, “Vote fraud! Vote fraud!” for the benefit of the camera. The Colliers were forced out of the room.
Even an average citizen should be a bit unsettled by the prospect of a single consortium providing all the data used by competing news organizations to discern winners and losers in national elections. To Kenneth F. Collier and his equally obsessed older brother James, the possibilities were apocalyptic.
In 1989, the brothers compiled the entirety of their research into 326 pages of manuscript—including a plethora of reprinted memos, clippings, court transcripts, and magazine articles. Their book is called, appropriately enough, Votescam. The ordinary person’s one chance to take part in democracy, the vote, has been stolen, says the book. Every significant election in the country, the Colliers believe, is fixed. And not by rogue opportunists or even Boss Tweed-style strong-arm “machines,” but by a sophisticated web of computer experts, media executives, and political operatives.
The brothers Collier, sons of a Royal Oak, Michigan, businessman, were both journalists. Jim had worked for the Miami News (though like so many impoverished reporters, he has already defected to public relations). Ken wrote features for the New York Daily News. In 1970 they caught the ear of an editor at Dell Publishing with a book proposal about running a grass-roots political campaign. The main chunk of research, they proposed, would consist of actually running such a campaign. And so Ken decided to take on the venerable Claude Pepper with Jim as his campaign manager and with no fundraising. The whole campaign cost $120 and consisted mainly of gumshoe canvassing, talking to nearly every voter in the eighteenth congressional district.
“It was a random thing that I happened to decide to run in the year 1970,” Ken told a radio interviewer in 1988. “But they had never used prognostications like this prior to that time in Florida. And when they did, it seems like we stumbled into the pilot project of the methodology that has since 1970 absolutely, completely, taken over the United States voting system.”
The Colliers’ revelation came on a date that lives in infamy for them alone: September 8, 1970, in Dade County, Florida. The events of that day appeared innocent enough. The Democratic party in Dade County held its primary election for the US House seat held by veteran congressman Claude Pepper. Pepper, who remained in Congress up to his death in 1989, was entrenched. He had no Republican opponent. The Democratic primary between Pepper and a hopelessly obscure opponent was de facto the final election, and a mere formality even in that regard.
The shock, to the Collier brothers, came soon after the polls closed at 7:00 PM. Two of Miami’s three television news stations projected Pepper the winner almost immediately. Nothing spectacular about that. They could have picked Pepper to win days before the election. What was remarkable were the exact predictions of Pepper’s victory margin and of the total voter turnout. At 7:24, one station projected a turnout less than 550 votes away from the eventual count of 96,499. In that same time span, less than half an hour, the stations called several other races on the ballot to within a percentage point of the final totals.
Unbelievable accuracy. But perhaps explainable as a marvel of technology, the genius of statisticians, or at least a mind-boggling stroke of luck. Until a University of Miami professor overseeing the projections announced one other fact: The projections were based on numbers from a single, computerized voting machine. Not one precinct, but one lone machine.
There was a third television station in Miami, but it was reported to suffer a computer malfunction on election night and waited until late in the evening to broadcast election results phoned in from county headquarters. By that time, televisions were off. Dade County received its results not from the courthouse, but from a single machine somewhere. Not even the professor who collected the spewing data knew where that machine was.
Most voters in Dade County watched the election returns with indifference. There were no big political surprises, least of all in the Claude Pepper race. The dazzling speed and precision of the local stations’ projections went largely ignored. Except, of course, by the Colliers, who were mortified.
According to the Colliers, the process used on a limited scale that evening in Miami has been expanded into an Olympian system that allows the three major television networks to “monolithically control” any election worth controlling—that is, most of them.
“What do they do? They wait ‘til the polls close. They announce who’s going to win in virtually every race, they announce what percentage these people are going to get. They are virtually never wrong. And the key to remember is once you have been named, you can rest assured you will be the winner. And later on, if only these networks can have some sort of mechanism whereby they could make the actual vote turn out the way they projected it nationwide, they would have the same setup they had down in Dade County, where they would announce who won early on, then meddle with the election results later to make sure they turned out that way.”