Some of the people behind 'wijvertrouwenstemcomputersniet' (we do not trust voting computers) are active in the CCC and have some pretty good arguments on why electronic voting really is a danger to democracy.
The link between electronic voting and online voting is a strong one and one would expect the online voting situation to be far more suspect to all kinds of trickery than the one where the voting computer is set up in a booth. Even so, there are some unique ways in which a voting computer in a booth might be manipulated to give incorrect results that do not apply to online voting, but I don't think it matters much, as soon as it's just bits & bytes and audit trails who watches the watchers becomes the real issue.
Any voting system without anonymity and a way to do a re-count and some physical proof is fundamentally broken.
I agree with some of the arguments against online voting. However, it isn't really realistic to do re-counts right now, nor do people do them regularly. If the people doing the counting are bought then that is an issue too.
Both ways suck right now.
One thing that I would like to see is an open source, standardized system though.
Stops Secrecy in Vote Tabulation: OVC has a team of scientists ready to program computer software for voting machines and electoral tabulation that would be publicly owned or open source. Open source software could be checked by any party or group by hiring a capable computer programmer.
Provides Paper Trail: The OVC recommended procedure for tabulating elections relies on a paper ballot that is then fed through a scanner into a locked ballot box so that all originals are saved in case of the need for a recount or audit (See Sample Ballot).
Scientifically Verifiable: In addition to open source voting machine and tabulation software, the Open Voting Consortium is also working on a database checklist for standard practices in vote tabulation that would assure transparency and accountability. Some aspects of the OVC concept will soon be enfolded into California legislation.
Saves Money: Typical voting machines cost between $2,000 and $3,000, but OVC open source software could be run on any personal computer (PC) and ballots could be printed on a normal printer. OVC envisions PCs with tamper-proof cases as the new voting terminals at a savings of hundreds or thousands of dollars per terminal.(See page on OVC Cost Analysis).
I don't get the complexity the US creates around voting. I live in a fairly large city (Berlin, Germany, around 3 million I believe).
- I walk 150m to my polling place. It's similar for most people.
- I've never waited more than 5 minutes
- The polling place is run mostly by volunteers, about 5 or 6 per place, with civil servants filling spots that can't be filled. Costs are neglible.
- I get a paper ballot, I mark my vote with a pen
- After 6pm I sometimes return and watch the counting. I can freely walk around between the people sorting the ballots and look at anything I want
- Did the same with friends last time around, we each had a bottle of beer, nobody cared. They were mostly older people happy to get a few visitors and talk a bit of politics.
- Pretty accurate first results (not exit polls) usually around 9pm.
- I can check the counts from my polling place the next day. Votes for the large party are in the lower 3-digits, so it's possible for me to specifically verify a few results.
Here's what most people that are not from the US don't understand:
When I go to vote on election day, I'm not voting for one person/party or even three people/parties. The typical ballot in my neck of the woods is double sided and has ~30 questions: National elections (President, HoR rep, maybe senator), state elections (Governor, SoS, local rep/senator, corporation commission, referendums/ballot initiatives), county elections (sheriff, judges), city elections (mayor, city council, dog catcher), and funding questions (bond overrides and the like). I'm probably forgetting an item or 10 on that list.
I never had 30 elections on one day in Germany (I think 5 was the absolute maximum) as we elect fewer positions in general and don't combine most elections on one day. Usually the federal election is the only vote on that day with some states specifically moving their state election day by one week to avoid a clash. The only usual combination is regional/European elections or state/regional elections for the small states. Referendums (there only a few) are usually tacked onto another election if there is one in a reasonable timeframe.
The elections each use a seperate ballot (and often seperate ballot box) and are counted one after the other with the most important election being counted first while the rest is still locked away in sight.
But yeah, for 30 ballots that system might break down.
I live in White Plains, NY, which is about 45 minutes north of NYC. My experience is similar to yours. But the US is a big, big place. I wouldn't be surprised if your experience has more to being in a major city, than in Germany. That is, what is the experience in rural parts of Germany? Poorer parts of Berlin?
Looking almost exactly the same. Rural (well rural in Germany is not the same as rural in the US) polling places might have 100 voters assigned instead of 1000 and people know each other but that's about it.
One issue is the cost and logistics of it all. In Sweden we therefore have three different elections on the same day. Campaigning tend to focus on the national election, thus the others gets a back seat and quality suffers.
Also, I've helped some smaller parties with the ballot distribution. It's not a small task for a small party to ensure each an every polling place in the country has ballots. Just the physical logistics of it is hard enough, but then you also have to deal with all the people who seem to think that only established parties has a right to participate in the election in the first place. The barrier to entry is quite high. Perhaps that is a feature though.
I have a question: Does every party has to supply a seperate piece of paper and the voter then chooses which piece of paper to put into the ballot box? What do voters do with the non-chosen ballots, i.e. how is the secrecy of the vote protected?
The system I know from Germany is that there is one ballot per election and the voter puts a mark which party/candidate they choose with a pen on that.
Yes. Three actually, if you count the different elections. On the ballot is a list of names representing the party so you can vote for a particular candidate by marking the name.
Usually people pick a set of ballots for different parties to bring behind a screen where the choosen ballots are put in the envelope for the box. The non-choosen ones you can pocket to keep secret. Some people tend to leave them in the booth though.
Voting procedures are handled at the state level, subject to terms from the federal government regarding basic access and fairness, so there is a lot of variety. I lived for twenty years in CA and I could have said the same thing about my polling experiences except for doing the counting (you can volunteer to be an elector and after some basic training do so, they just don't let the general public wander in to the best of my knowledge.)
In France, if you can vote, you can even count, and people can volunteer when they cast their ballot. The president of the voting station will usually also ask young voters to help.
> However, it isn't really realistic to do re-counts right now, nor do people do them regularly.
Australia uses paper ballots, and every election there's a seat or two somewhere that there's a recount. The major parties get their volunteers[0] to scrutinise the officials as they count as well, so you have multiple opposing interests scrutinising the actual count. Every vote basically gets at least three pairs of eyeballs on it. The officials and the watchers observe the ballot boxes all the way from the booths to the counting areas. It's difficult to come up with a more robust way of doing an actual public election, though there are some minor drawbacks.
And as far as I'm aware, this system is actually cheaper (per capita) than the US electronic systems. I imagine the 'low tech' nature is where a lot of savings happen - little need for technical skills.
[0]'Scrutineering' is boring as hell, but parties already lean heavily on their volunteers, what's one more thing?
Italy here, it works in the same way. Parties have their watchers. I expect this to happen in every country, unless there is only one real party and the others are there just for the show.
In every paper based system problems can arise outside the voting site: parties can buy votes but that's the same with internet voting. With the internet and computers an attacker has the benefit of changing a vote every 1,000, on the client or on the server or on the network (MITM). That's enough to win many close elections. I wonder how parties could watch against that. Are we going to end up with computers with rootkits for every major party fighting each other for the right to vote on our behalf?
The paper based system guards against vote buying through the guarantee of secrecy. I go into a booth on my own, make my mark(s) fold my paper, and put it in a locked box. No one but me knows how I voted.
If I go in with someone else, the officials will take action.
You can try to buy my vote, but I can just take your money and vote however I want.
If I can vote from wherever I like, then someone can stand over my shoulder and watch me vote, then only pay for it if I do as I'm told.
Sending a vote buyer knocking on doors and watching people vote when they happen to be voting is not very scalable. However, party activists and pollsters already do the first part. Adding the second isn't a massive leap.
However, there are situations that might scale better.
A business owner or manager might strongly persuade their workforce into voting the "right" way, by getting them to vote at work. This would scale quite well in a non-office environment, where you might suggest to your workers to come and use one of the few computers you have (for their own convenience, of course).
Another option might be some kind of party, where you invite people to come and vote for your candidate in exchange for payment. This could work quite well to gain the votes of the poorest in society, who may otherwise find it difficult to vote (no internet at home, can't easily get to the polling station).
Remember, those who care enough and are decisive enough to vote early in the day, are not the target of vote buyers. The people whose votes you can buy are those who wee possibly thinking of not voting at all, or those yet to decide as the polls are closing.
Think about the scalability of the other side. At the moment, you have a few people per few thousand voters sitting in the designated room for those voters, watching out for violations.
Both of these intimidation/buying activities would be illegal, and would certainly not be endorsed by the actual party in question, but by rogue individual supporters of the party. However, detection and prosecution of offenders will be far more difficult and time consuming.
Having to detect and prosecute these irregularities after-the-fact would also mean that elections may have to be invalidated and redone (possibly ad-infinitum) when a violation has happened, if you just remove all the illegal votes and recount, then what about all those voters who did vote with their conscience, but just happened to do so in the wrong place. Then what do you do about the actions of the illegally elected in the interim?
Although, calling this system a "Paper ballot" is not quite the truth. While it uses paper as the medium, the hole punching method is non-traditional in relation to basic written checks/X's or even scantron selection (which also has issues). But my point is that even "simple" and "low tech" methods can still have unexpected issues, such as the infamous Florida case.
> However, it isn't really realistic to do re-counts right now, nor do people do them regularly.
About half of the states require by law a recount by hand at a random selection of voting locations.
> If the people doing the counting are bought then that is an issue too.
That's why you have hostile observers from opposing parties present.
> open source software
So you have the source. Now prove that it is installed at both the client and server. Then prove that the operating system, bootloader, and every other part of the computer you need to trust is actually loading and running your software properly.
Difficulty: Volkswagen showed it's easy evade inspections, and good luck proving anything on a system with Intel ME. Do you know that there isn't an extra core or ROM hidden inside one of the chips?
> > If the people doing the counting are bought then that is an issue too.
> That's why you have hostile observers from opposing parties present.
That's probably why GP wrote "bought", not "asked nicely to miscount in their party's favour".
That said, it's probably a game neither party would like to play too much, as it seems it could spiral out of control and turn into a race to bottom, sucking out more and more party money.
>> That's why you have hostile observers from opposing parties present.
Even though observers are supposed to behave in an impartial and accurate way, partisan observers help ensure accuracy by their partisanship.
If an individual counter is bought by Party A, observers from Parties B and C are there. When the teller puts some of Party C's votes in Party A's pile, Party C is there to bring it to the attention of officials. If it is likely to be close between A and B, and a teller puts a spoiled, party D, or independent paper in A's pile, then B will note it.
Once these things have been raised, they are recorded and read by more officials and observers. That increases the number of people you need to buy off in order to effect a change in the result, which in turn increases the chance of being caught or having to buy off someone who cannot be bought in your favour.
Fundamentally you could buy people off. However you'd need to get the people counting, and the officials, and people in different parties, and not be spotted by any of the public that watch. And then do that again, simultaneously, in thousands of other locations, and not get caught or have anyone you tried to buy out go to the police.
San Francisco has started the process of creating just that kind of end-to-end open source voting system (using the existing paper ballots, optical scanning and manual recount process that exists here). Here's some more information: https://gcn.com/articles/2016/06/02/sf-open-source-voting.as...
Not really, you just have to be given a number that you can later match, no one has to know that number is you. Could even digitally sign things.
I propose a system where each vote is broadcast to multiple counting organisations, to avoid the counters being biased, and each vote is numbered, but only the voter knows that number so they can check with any counting organisation that it was recorded as they wished.
E-voting makes better forms of voting cheaper, instead of just First Past The Post. Preference voting for example.
If you just give out a number you suddenly open up the system to voter coercion. I can actually prove to you how I voted. Also, as a voter I can't be sure that they system that gives me the number does not save the link between it and me.
For preference voting you don't need e-voting. Using computers to help you count ballots (with an analog check and fallback option) is possible without actually collecting the votes electronically.
Much more of an issue that people don't vote than people sell their votes, which they do indirectly anyway.
Does anywhere use computer printed ballots?, to avoid ambiguously filled out ballots.
Computer creates PNG, you print that, or it's printed at a station, you carry that paper to the vote counting machine, it's optically read, (and read by humans if you wish) without errors because it was just printed by a computer
Which is why the voting machine should print a paper receipt which in plain text tells what vote you cast; this receipt is then deposited in a ballot box.
If a result is disputed, simply count the paper ballots like we do today.
Unless the result is contested you have saved an amount of counting effort which could be a cost saving. Though I suspect it would take some time for the ROI of implementing such a system to work out positive.
> > Unless the result is contested
> > you have saved an amount of counting effort
> The ballots are already optically scanned.
If the result is contested then I presume that this scanning may also be under suspicion (of being faulty or having been somehow doctored deliberately).
> > cost savings
> A handful of pencils per-voting-booth is far cheaper than any computer+printer solution.
Why is why I mentioned the time for ROI to become positive. The electronic counting will save time (man power for the initial count, the need for recounts, and just wall-clock time so the results are available earlier) but the cost will be high. At some point perhaps the solution will pay for itself in time/resources saved but it will take quite a while if at all.
Maybe you should study how elections are actually implemented instead of making assumptions. I suggest watching the talk I linked to in my top-level post.
> this scanning may also be under suspicion
Obviously. Hence why it's important to recount the paper ballots by hand. This is also why about half of the states require confirming the reliability of the optical scanners with a random sampling of hand counts.
> At some point perhaps the solution will pay for itself
No, it won't. Computers per-voting-booth are always going to be a lot more expensive than an optical scan tabulator per-site.
Also, because a computer+printer solution involves a lot more devices, you will need a larger amount of testing by hand recount. You are increasing the workload.
And since you can't know that the computer actually recorded what was printed on your paper receipt you have to dispute the result by default if a good democratic process is important to you.
So if you have to count the paper balloty anyway, why even bother with spending all that money for a voting machine?
There's at least two reasons, possibly even three:
a) Unless the outcome balances on a handful of votes, chances are noone will contest it. (If the result shows 47% to Party A, 32% to Party B and 21% to Party C, whereas polls showed 46% to Party A, 33% to Party B and 21% to Party C, chances are the result is fair.
That being said, yes - I think it would be good if just about anyone could contest the result and have a recount done - at least if they had to foot part of the bill, so that not every curmudgeon in the Kingdom would claim recounts just for the hell of it.
b) A voting machine system would enable one to have a preliminary result ready in seconds after ballot stations close, rather than today's system where, based on the size of electoral districts, results may be delayed by several days.
c) If properly designed, a voting machine could assist the user in creating a valid ballot; I've volunteered as an electoral clerk several times - we have had to dismiss surprisingly many ballots as there is simply no (approved) way to determine voter intent. A few checks before the vote is cast would likely improve matters.
https://www.ccc.de/en/tags/wahlcomputer
The link between electronic voting and online voting is a strong one and one would expect the online voting situation to be far more suspect to all kinds of trickery than the one where the voting computer is set up in a booth. Even so, there are some unique ways in which a voting computer in a booth might be manipulated to give incorrect results that do not apply to online voting, but I don't think it matters much, as soon as it's just bits & bytes and audit trails who watches the watchers becomes the real issue.
Any voting system without anonymity and a way to do a re-count and some physical proof is fundamentally broken.