Electronic Voting – at what cost?



In the year 2000, the United States Presidential Election was decided by the Supreme Court due to electoral irregularities in the state of Florida. Many voters who intended to cast a vote for Democratic candidate Al Gore had mistakenly cast a vote for another candidate.

In the aftermath of the 2000 election, electronic voting was identified as a possible solution to the problems that occurred, although new issues were raised regarding the security of ballots, with concerns that an election could be hacked or that faulty software code could result in an incorrect count.

Electronic voting was instituted in some regions for the following electoral cycle, with faults and irregularities being widely reported with the new electronic voting machines.

In Australia, electronic voting has been used in a limited way since October 2001, with the Australian Capital Territory being the only electoral jurisdiction that has incorporated electronic voting with its normal election processes. In the ACT, electronic voting terminal are linked to a server in each polling place using a secure local area network, with votes transferred to the counting centre on a compact disk. No votes are taken or transmitted over a public network such as the Internet. These systems provide voting facilities for vision impaired voters as well as able-bodied voters. This facility has provided the opportunity for vision impaired voters to cast a secret ballot.

According the NSW Electoral Commission’s (2010) report a number of electoral jurisdictions have conducted electronic voting trials. Including:

• The Australian Electoral Commission (AEC) trials for electronic voting in 2007

• Victoria piloted electronic voting machines for vision impaired and blind voters in 2006

• Tasmania made electronic voting machines available for vision impaired and blind voters in 2007

Some of the lessons learned from the trials were:

• The cost per vote for these systems can be high, if the adoption rate is low.

• Feedback collected from these trials indicate that the term “electronic” has discouraged a significant number of older, less computer-literate voters

The Joint Standing Committee on Electoral Matters (2009) in Australia, states on the Report on the 2007 Federal Election Electronic Voting trials that to cast a vote electronically it costs $2,597 per vote, while on average the cost per elector in regular federal elections in 2007 is $8.36. The report also says that due to the low participation in the trial it appears to be unsustainable and hence the committee does not consider that electronically assisted voting to be a permanent feature of federal elections.

If mainstream voting is to go electronic, systems must be made available on a high-availability basis, in situations and locations that are not designed for such demands. If electronic voting suffers a localised or widespread outage, what sort of backups should be in place?

Electronic voting should not necessarily mean voting over the Internet. One of the safeguards of our secret ballot is that voters must present themselves at a polling place, and for the most part, be separate from other people while they fill in their ballot papers. If voting were to take place over the Internet, it would not take long for union thugs to setup their own polling place in factories, construction sites, school staff rooms, government agencies and other industries where trade unions still have influence. One could envisage a situation where workers are shunted into the union-provided polling place, with a union delegate standing over workers and instructing them that “you’re voting Labor”.

An electronic voting system could be run in a secure manner, or in a more secure manner than many paper ballot methods that are currently in widespread use. One such model involves voters still turning up to polling places run by agencies such as the Australian Electoral Commission or their state equivalents. Voting machines with publicly released machine code would provide voters an on-screen choice of candidates. Voters would make their selection, and the voting machine would then print a completed ballot which the voter would then look at to ensure that the printed ballot is as they intended. If all is well, the voter would then place the printed ballot into a ballot box. If the printed ballot paper is not as intended, the voter could speak to a voting official and have that ballot paper invalidated and destroyed, and be allowed to cast a new ballot.

Reasonable security precautions could be built into such a system, such as voters being issued a barcode or a Radio Frequency Identification (RFID) tag to enable them to vote, and for printed ballot papers featuring a barcode so that a vote could be invalidated on an error having been made. The barcode or RFID tag issued to the voter must not have any correlation or link to the end ballot in order to ensure the vote remains secret. These barcodes or RFID tags should have further security measures, such as being time limited to a period of around 5 to 10 minutes and only valid at a particular polling place, to reduce the probability that tokens could be passed off to another person to cast many votes, and to protect against the union influence scenario.

At the close of polling, votes would be counted by the computers, and a further count by separate and disconnected machines, of the paper ballots that were printed and placed into the ballot box. If the counts match – great. If the counts do not match, votes cast at that polling booth would warrant further scrutiny, with the printed paper ballots being used for the official count as they would have been visually inspected by voters.

Other matters would still need to be considered, such as the transmission of votes from the polling place to a centralised location. Encryption and Virtual Private Networks (VPN) could be used over the public Internet, or individual results could be called in by the official in charge at the polling booth – or a combination.

Electronic voting can work, although I have significant concerns with there being a widespread Internet based voting system for end users/voters. It is significantly expensive at this stage, and until a cost benefit analysis shows that it can be implemented at a cost more comparable to traditional voting, and while incidents of voter fraud within the traditional paper based system is very low, electronic voting should not be implemented, although research and development in this area should take place in order to reduce costs, to achieve a more accurate count and to obtain results in a much shorter timeframe.


Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s