Wednesday, October 19, 2005

'E-Stonia' online vote called a success

The first reports on the Internet voting experiment in Estonia are coming in (first blogged about by Alvarez and Hall here, and by all reports the system was a success.

There is no statement in the story about how many people the Estonian authorities hoped would use the system, but reports are that approximately 1% of the population voted over the Internet.

The Estonian system implemented an inexpensive ID card system--the card reader costs $24 and each individual was issued a unique ID card. About 80% of Estonians already have the cards, according to official reports, because they are already used for access to tax records and bank accounts. (Ed: not sure what that means--does Estonia have a nationalized banking system?)

This system appears to be effective and relatively inexpensive. I commented on a similar sort of security system in a recent interview I gave to the WSJ on telephone and internet voting. I described thumbprint recognition systems which, although they sound high tech, are actually quite inexpensive in this day and age.

Jason Kitcat (unfortunate name for anyone familiar with US candy) criticized the system, but implicitly criticizes any computer based voting system. I'm not sure that is a realistic position in this day and age.

Alvarez and Hall have written a lot more about this in the recent Brookings volume. Perhaps they can post a bit more about the security protocol used in Estonia.

Story referenced is here.

3 Comments:

Blogger Jason Kitcat said...

I'm not sure on what basis you can call the Estonian ID card system inexpensive... The cost of rolling out the cards, PKI infrastructure etc was considerable (I don't have precise numbers). Furthermore $24 for a card reader is rather a lot, particularly considering the low average salaries in Eastern Europe.

Thumbprint recognition systems are fairly cheap but not workable on national scales due to their high error rates. If you're using it to login to your computer then it only has to check the thumb presented against one record. When you have 60, 100 million thumbs false positives and negatives become prohibitive.

Opposition to e-voting is a perfectly realistic position today. Apart from in the US where things have progressed too far. Technology is great but the few benefits it offers to the voting process do not out the huge risks it poses. If we can weigh up the risks and say 'no' what's wrong with that? Technology isn't always inevitable.

(By the way, KitKat is a British chocolate bar dating back to 1935 see more on Wikipedia, my name and the chocolate's are intertwined and probably all go back to the Kit Cat Club)

12:55 PM  
Blogger Paul Gronke said...

My apologies about the candy reference. Too cute and too wrong.

I don't think we agree on thumbprint recognition--my presumption would be that the system would operate at the individual computer level, verifying the validated user--not against a national repository.

Even if there were a national repository (something inconceivable in the US), the system would not check your thumbprint against all existing prints--it would only verify it for your registration record.

As to the cost of the cards, I don't expect the individual to pay for these. In Estonia, my understanding is that users bought the cards for use with other linked systems. But I have no problem having the government shoulder this cost. At $24 per person, this would be a substantial savings over other election administration systems.

9:43 PM  
Blogger Jason Kitcat said...

Your presumption that it wouldn't be a national system may be correct in the US. But the UK is actually trying to build a national database of just those kinds of biometrics.

The Estonian cards are from a national identity program, as there is across much of the EU. However they have an integral, and useful, public key infrastructure with the require accompanying legislation. Thus signing a transaction with the cryptographic key on the card is legally accepted in many transactions including buying a house.

If $24 was the only cost and the readers worked universally (they don't) and voters didn't lose them between elections then perhaps there would be a cost saving. But e-voting delivering cost savings is a mirage which even the vendors don't persist any more. e-voting is more expensive than traditional voting, the data shows it, the vendors don't deny it. It's expensive and risky.

More on e-voting's pros and cons at my site

2:45 AM  

Post a Comment

<< Home