How many legit votes prevented for each fraudulent vote?

Great Debates
1

All of the voting threads have got me wondering: Most measures to prevent vote fraud increase the difficulty of voting, convincing some people who would have voted to skip the hassle. In the absence of an obviously disproportionate impact on one constituency, how much hassle is preventing a fraudulent vote worth?

It seems obvious to me that a lower bound would be one legit to one fraudulent, because a fraudulent vote can cancel out a legit vote. In cases in which fraud is one-sided but the ballot issue is nearly even, the ability to stuff 100 ballots may outweigh 10,000 additional votes, 5050 for and 4950 against. Clearly ballot-box security is important enough to invest in some inconveniencing measures (perhaps a line with a couple observers per box, or ballots limited to one per person and tightly controlled). I provisionally accept 100 legit votes discouraged per fraudulent vote prevented, but will probably settle on a lower value.

But higher numbers? I understand the sentiment that the least motivated voters are less important, but I don’t think I could get behind measures that discourage 1000 people out of voting to prevent one fraudulent vote. Is the reasoning based on a principle that it is infinitely more important to prevent fraud than to have accessible voting?

To estimate my line of thinking- say you live an a city of 1 million people. If this city has unusually high turnout, you may get 500,000 votes (60% turnout * 80% voting age). A very unusually close election (perhaps fantastically so) may be decided by 1,000 votes. To prevent systematic fraud from reversing the election, would you take measures discouraging 1000 votes? 10,000 votes? Would anyone claim that the election is more valid if security is so rigorous that 100,000 or more votes are discouraged? It seems to me that beyond the simplest measures already present, tighter controls make elections less trustworthy, not more.

I understand that it is difficult to estimate the amount of legitimate votes discouraged, and even harder to estimate the number of fraudulent votes prevented. If we could know, where would you draw the line, and why?

I also understand that discouraged voters will not be drawn proportionally from each constituency. It is my view that measures that have a disproportionate impact on different constituencies should have an even lower threshold for prevented legit vs. fraudulent votes, but I’m not sure how much lower. If possible, I’d like to at least start this debate in hypothetical land, where security measures do not have a disproportionate impact on the electorate and are implemented solely to prevent fraudulent votes.

2

I think the principle should be denied votes and fraudulent votes should be combined together and we should try to get the total as low as possible. A “tight” system that denied 999 legitimate votes and allowed one fraudulent vote would score 1000. A “loose” system that allowed 999 fraudulent votes and denied only one legitimate vote would also score 1000. A medium system that prevented 400 fraudulent votes and denied 400 legitimate votes would score 800 and would, in my opinion, be better than either of the other two alternatives.

3

I don’t see any hassle involved in registering to vote. It’s not a complex process in the least. In fact, I can’t think of one single process that is easier to do in society than registering to vote. Every other form of registry involves a great deal of paperwork.

4

I don’t think the OP was necessarily saying that any current real-world voting requirements were too strict; but hypothetically, one could set up an extremely rigorous registration/identification process.

For example, in Canada to get a passport you used to have to find a guarenteer that a) knew you personally for 5+ years b) was a member of a certain profession (doctor, lawyer, engineer, accountant, etc.) and c) was will to sign your application and photographs as well as provided contact information to the passport office. On one of my applications they even called up my guarenteer and quizzed her about me.

If one really, really wanted to be extra sure that every voter was legitimate one could create a ‘voter ID’ that is even harder to get than a passport under the old Canadian system. You could be required to have 3 guarenteers, and they would have to know you personally for 10+ years, and all 4 of you have to apply in person, at the same time. This is of course an insane example but in principle it’s the same for more realistic systems. The more strict the requirements, the fewer potential voters there are that can meet them or are willing to jump through the hoops to do so.

In answer to the OP if the numbers could actually be measured I think I’d accept something like 3:1 discouraged to fraudulent votes.

5

Minus a method of verifying that voters are legal voters, there’s no way to rate any proposal. If you do verify that voters are legal, then you can no longer make the comparison.

6

Wisconsin study showed 20 percent of the voters in last election were without picture ID.

7

Do you have a link to that Wisconsin study?