Vote Counting, Technology, and Unintended Consequences

Abstract:     
The 2000 presidential election had a searing effect on this nation. Few who witnessed the events in Florida could displace the images of election officials peering at punch cards, struggling to determine the intent of voters. Congress, for example, did not forget. Congress did not wish to see the scenes from Florida replayed in future elections. And so, in 2002, it enacted the Help America Vote Act, known as HAVA, which provided $325 million to the states to replace their punch card voting systems.

Many states have enthusiastically embraced this invitation, replacing punch cards with electronic voting machines, known as direct recording electronic devices (DREs). Given the rapidly approaching 2006 deadline to upgrade, other states are currently considering such activity.

In the rush to solve one problem, however, states may be creating another, far greater, one. To be sure, DREs promise to reduce certain errors, increase access for voters with disabilities, and relieve election officials of the challenge of ascertaining the voters' intent.

But the full panoply of dangers from such systems have largely avoided scrutiny. This Article attempts to remedy this deficiency. In particular, it underscores several disturbing characteristics of electronic voting, including reduced transparency, increased magnitude of error and fraud, and lack of security controls.

I supplement the analysis of the DRE software by examining vote counting flaws in the 2004 presidential election, including machine breakdowns, vote totals that exceeded or underrepresented the number of voters who cast ballots, and incidents in which the machines switched votes from one candidate to another. I also collect circumstantial evidence such as exit polls and pre-election polls that significantly diverged from the official vote count and questionable activity in states such as Ohio. Although such evidence does not automatically prove the existence of error or fraud, it is crucial given the nontransparent nature of the vote counting process and inability to directly uncover fraud.

I conclude by offering recommendations to improve the accuracy and verifiability of vote counting today. In particular, I propose for electronic voting machines a voter-verified paper trail, random audits, open source software, and other recommendations. Only after these proposals are adopted can voters have confidence that the promise of vote counting technology will match its perils.

Keywords: Electronic voting, vote counting, computerization, technology, HAVA, 2004 election

JEL Classifications: D7, D72

Accepted Paper Series