Like
it or not, computer technology will be used for most elections in
some way or other before too long. Already, even in impoverished
countries like Nicaragua, centralized electoral systems use computers
to manage the counting process. Brazil votes using computerized
systems. In the United States, people worried about electoral fraud
are becoming more vocal—with good reason. The Help America
Vote Act (HAVA), passed in 2000, and the Omnibus Appropriations
Bill, approved by President Bush in October 2002, will force electoral
authorities throughout the U.S. to adopt computerized voting systems
by January 2006.
But
the record of the leading companies supplying computerized electoral
systems in the United States is questionable. Companies like Diebold,
ES&S, and Sequoia— who have over 80 percent of the market
for automated voting systems in the U.S.—have all been criticized
for having problems with their machines. Apart from software and
machinery malfunctions, computerized databases of voters also cause
problems, as when the Choicepoint data systems company subsidiary,
DBT, incorrectly purg- ed over 90,000 registered voters from Florida
electoral lists in the 2000 presidential elections. Additionally,
the use of modems leaves the way open for data to be tampered with.
Three
main points of view prevail on these issues. Proponents of computerized
voting systems argue the need to modernize so as to facilitate voting
and draw more people into the electoral process. Other advocates,
such as Public Citizen’s Congress Watch, say of HAVA, “In
many ways, the new law marks a significant step forward in improving
the conduct of elections in the United States. At the same time,
however, the compromise sacrificed some additional steps that should
have been taken to ensure that every vote counts and contains some
of the ballot security measures that are not useful to the democratic
process.”
Critics
disagree strongly. Rebecca Mercuri, a computer science professor
at Bryn Mawr College in Pennsylvania, has researched computerized
voting for over a decade. She asserts, “Fully electronic systems
do not provide any way that the voter can truly verify that the
ballot cast corresponds to that being recorded, transmitted, or
tabulated. Any programmer can write code that displays one thing
on a screen, records something else, and prints yet another result.
There is no known way to ensure that this is not happening inside
of a voting system.” Mercuri and other critics point out that
electronic balloting systems without individual print-outs for examination
by the voters do not provide an independent audit trail. They also
question the lack of certification to international computer security
standards of electronic voting systems. Another main concern with
these systems is the shift of control away from election officials
to computer personnel.
Peter
G. Neumann of Risks Forum, who monitors problems with computer technology,
writes, “We have reported election problems in Software Engineering
Notes and Risks for many years…. We note that punched-card systems
are inherently flaky, and that even optical scanning is problematic,
but that direct-recording electronic systems tend to be subject
to serious potentials for fraud and manipulation. Internet voting
is a disaster waiting to happen in light of the inadequate security
of the Internet, personal computer systems, and subvertable servers.
Proposals to vote from automated teller machines are…basically
undesirable.”
Proponents
of computerized voting systems—often owned by large transnational
businesses—argue that security is good and machines conform
to government standards. For example the ES&S company web- site
states: “ES&S products are tested by an independent testing
authority, certified to meet or exceed the standards of the U.S.
Federal Election Commission and have been proven and validated through
use in thousands of actual elections worldwide.” Peter Neumann
responds to these assertions: “The Federal Election Commission
standards that are in general use appear to be those from the 1990s.
The review process that was used for the revised 2002 standards
was seriously flawed and many of the review comments were ignored
almost completely. As a result, the newly approved revised standards
are fundamentally inadequate.”
To
adjudicate these competing claims a look at real world experience
may help. In August 2002, the results of at least 18 suburban Dallas
County elections were delayed through vote-counting problems using
ES&S software. The Dallas Morning News report on the
glitch referred to “Election Systems & Software, the company
that sold the previously trouble-free equipment to the county four
years ago.”
Trouble
free? Here’s what the Venezuelan national electoral authority
had to say about ES&S in May 2000. “We say ES&S has
not been sufficiently efficient in testing what it was supposed
to have supplied…the National Electoral Council cannot accept
such a failure of responsibility by this North American company.”
So the Venezuelan elections scheduled for May 28 that year were
cancelled. In November 1998, faulty ES&S voting machines used
in Hawaii on election day “led to Hawaii’s first ever
statewide election review and a first in the history of the United
States.”
The
practical problems with electronic voting are well documented. Peter
Neumann’s Risks Forum posts a daunting list of errors and failures.
The recent highly publicized case of the Diebold company’s
electoral software, discovered by chance on an open website, downloaded
and tested by computer scientists, confirms that fears with security
and use of databases and software passwords are justified.
But
apart from the technical aspects, computer voting raises old issues
of undue influence and interference in a new guise. In 1999, 22
people were indicted in Louisiana (9 admitted guilt) in a huge bribery
scam involving the acquisition of Sequoia voting systems. Sequoia
Pacific’s Regional Manager and a regional sales executive were
indicted for paying around $8 million in bribes to Louisiana Commissioner
of Elections Jerry Fowler. Fowler was sentenced to over four years
in prison.
In
Georgia companies are vying for a $54 million contract to supply
18,000 touch screen voting machines. Among the competing contractors
are Diebold Election Systems and Northrop Grumann Diversified Dynamics.
Diebold’s CEO, Wally O’Dell, is a major fundraiser for
the Ohio Republican Party. (Ohio Democratic leaders are seeking
to block Diebold’s bid to supply voting machines to the state.)
Northrop Grumman Corporation is a major defense contractor with
links to the Carlyle business group, a nest of eminent Republicans,
including former President George Bush.
The
involvement of big business in the management of electoral databases
and computing of votes is inherently and profoundly undemocratic.
Politicians have come to see manipulation of the vote much as they
see gerrymandering boundaries of voting districts—all part
of the electoral game. For many people disenchanted with politics
and politicians, the system has long been not one person/one vote
but one dollar/one vote.
The
case of Senator Chuck Hagel exemplifies concerns about questionable
business and political links. In 1996, Hagel won a totally unexpected
victory in an election where his own company’s computerized
voting systems was doing the counting. He was the first Republican
in 24 years to make the Senate in Nebraska. In 2002, Hagel ran again
and was elected with 83 percent of the vote. Thom Hartmann observes,
“80 percent of those votes were counted by computer-controlled
voting machines put in place by the company affiliated with Hagel:
built by that company; programmed by that com- pany; chips supplied
by that company.”
Against-the-trend
results swung the Senate for the Republicans in the 2002 elections.
In Georgia popular Democrat Max Cleland was leading the pre-election
polls 49 percent to 44 percent. Mysteriously, his lead evaporated
on election day, turning into a 53 percent to 46 percent win for
his opponent Saxby Chambliss. In Georgia, Democrat Roy Barnes led
Republican Sonny Perdue in the opinion polls by 48 percent to 39
percent. Nonetheless, Perdue won with 52 percent of the vote against
Barnes’s 45 percent. In Minnesota, just days before the election,
veteran Democrat Walter Mondale—a late replacement after the
death in a plane crash of leading Democrat Senator Paul Wellstone—led
Republican Norm Coleman by 47 percent to 39 percent in opinion polls.
But Coleman won, 50 percent to 47 percent. In all these states computerized
voting systems were used to count most of the votes. It seems very
strange, to say the least, that opinion polls in three states should
have goofed so badly.
Foreign
involvement in those companies is another issue. Sequoia is owned
by De La Rue, the British security systems transnational with a
minority shareholding by the Dublin-based Jefferson Smurfit Group,
another transnational company. A contract to record the votes of
the U.S. military has been awarded to Accenture, a Bermuda-based
company formerly part of the Andersen auditing group, so thoroughly
discredited during the Enron collapse. These transnationals work
comfortably with the business interests currently running the White
House.
The
available evidence indicates that the 2004 election could be spectacularly
and, in most cases, undetectably rigged using computerized systems
supplied and managed by companies linked to the Republican party.
New
technology—vulnerable to tampering—will certainly be put
in place under HAVA. The resulting mess will be adjudicated in the
courts, if disputed results ever get that far. Four protections
are needed to prevent voting manipulations:
effective monitoring of databases to prevent purging of legitimate
voters;
a physical audit trail so people can be sure not only that their
vote is registered correctly, but that someone can verify it;
strong, legally enforceable statutory standards for all computerized
voting systems and voter database systems (not contemplated
in HAVA, which empowers Electoral Standards Boards to implement
only vague “voluntary guidelines”);
open, non-proprietary verifiable software for all these systems.
A
recent report from Johns Hopkins University on computerized voting
systems concluded: “…there is little difference in the way
code is developed for voting machines relative to other commercial
endeavors. In fact, we believe that an open process would result
in more careful development as more scientists, software engineers,
political activists, and others who value their democracy would
be paying attention to the quality of the software used for their
elections…such open design processes have proven very useful in
projects ranging from very focused efforts…through very large
and complex systems such as maintaining the Linux operating system.”
Phil
Hughes runs the World- Watch website, which addresses social, political,
and economic aspects of Linux and non-proprietary open- source software.
He explains, “Linux is a free and open operating system standard
developed by people all around the world. Linux and many applications
programs written for Linux are produced under a public license agreement.
No one has a monopoly of the basic product information…
“It
would take about six months to develop viable free open source electoral
software in Linux. Obviously it would require testing. But much
existing commercially developed electoral software is still showing
problems despite many years of use. Linux will certainly do better
and be more reliable because it is open source, available to everyone.”
Free,
open to everyone? Sounds just like what the computerized voting
market needs, and fast.
Toni Solo is
an activist living in Central America.
