Increasing voter turnout with compulsory voting and (gasp) electronic voting

Earlier this year, I started a series on fixing U.S. democracy. Today let me look at the problem I identified as #3: Voter turnout and the excessive power of GOTV.

In a big political campaign, fundraising is king, and most of the money goes to broadcast advertising. But a lot of that advertising, a lot of the other money, and most of the volunteer effort goes to something else called GOTV or "Get Out the Vote." Come to help a campaign and it's likely that's what you will be asked to do.

US elections have terrible turnout. Under 50% in the 1996 Presidential election, and only 57% in more recent contested elections. In off-years and local elections, the turnout is astonishingly low. Turnout is very low in certain minorities as well.

Because turnout is so low, the most cost effective way to gain a vote for your side is to convince somebody who weakly supports you to show up at the polls on election day. Your ads may pretend to attempt to sway people from the other side, or the small number of "undecideds," but a large fraction of the ads are just trying to make sure your supporters take the trouble to vote. Most of them won't, but those you can get count as much as any other vote you get. So you visit and phone all these mild supporters, you offer them rides to the polling place, you do everything legal you can to identify them and get them out, and in some cases, to scare the supporters of your opponent.

Is this how a nation should elect its leaders? By who can do the best job at getting the lukewarm supporters to make the trip on election day? It seems wrong. I will go even further, and suggest that the 45% or more who don't vote are in some sense "disenfranchised." Clearly not in the strong sense of that word, where we talk about voter suppression or legal battles. But something about the political system has made them feel it is too much of a burden to vote and so they don't. Those who do care find that hard to credit, they think of them as just lazy, or apathetic, and wonder if we really want to hear the voice of such people.

GOTV costs money, and as such, it is a large factor in what corrupts our politics. If GOTV becomes less effective, it can help reduce the influence of money in politics. It's serious work. Many campaigns send out people to canvass the neighbourhoods not to try to sway you, but just to figure out who is worth working on for GOTV.

Compulsory voting

Many countries in the world make it compulsory to vote. If your name is not checked off at the polling place, you get fined. Australia is often given as an example of this, with a 91% turnout, though countries like Austria and New Zealand do better without compulsory voting. But it does seem to make a difference.

Compulsory voting would severely reduce GOTV if the penalty were high enough. If combined with making voting easier, it could come close to killing GOTV's power. It should also be pointed out that a person who votes for this reason should have the right to submit a blank ballot (usually they already can) and that if voting is made very easy, they can even be encouraged to submit a blank ballot if they truly don't care. But anybody who cares mildly -- the targets of GOTV -- would just vote if it were easy and compulsory.

You may feel we don't want these lukewarm voters to be deciding our elections, but they already are, or worse, whoever has the bigger GOTV machine is what decides our elections by exploiting them. As much as we might like it, "Elections should be decided only by those who care enough to make the effort" just isn't an option on the table.

Compulsory voting would probably have to be enacted by individual states, not federally under the constitution. Some argue the constitution still forbids this -- that you have a right not to vote as much as you have a right to vote -- but this might be addressed by the explicit blank ballot. And Obamacare recently showed you can kludge around this with a tax rather than a fine.

Easy voting at home

The states with the highest turnout make it easy to vote. Oregon has all vote-by-mail and has the 6th highest turnout. Minnesota is the consistent winner with 67% turnout due to same-day registration, as does Maine at #2. But Oregon's 60%, while 8% above the national average, is hardly a great victory.

This has made me reconsider something I've had a longstanding opposition to, namely electronic voting, and beyond that, electronic voting at home.

Electronic Voting

We've been fighting electronic voting for a long time because it's very hard to make it secure. Most voting machine designs have been terrible, and create a real risk that somebody could corrupt them to alter elections in a way we couldn't detect by auditing. There are much better designs used in a few places, but the first wave of machines were clearly flawed and had to be fought.

Electronic voting at home has always been a non-starter. Your home computer, your phone and the public data networks are insecure. They can be easily hacked and made to steal or corrupt your vote if you used them. Electronic voting at home also must be counted in a purely electronic central computer which can be corrupted.

But electronic voting offers a number of advantages. US elections are hugely complex compared to the rest of the world, with dozens of races on typical ballots. A nice screen based UI can guide voters through the process, and work in many languages or even with voice for the illiterate. The millions of disabled, notably the blind, also feel voting machines are a blessing, allowing them the fundamental right to vote in private for the first time.

Voting from home would offer huge benefits as well. It eliminates many types of voter suppression around polling places, or people who don't have enough time to vote on a workday. It eliminates the hour-long waits found at many polls in busy elections which clearly disenfranchise or discourage many. The difference is so huge, I've decided to give electronic voting a second look.

Hybrid paper/electric voting

Ten years ago, I designed a system for secure hybrid electronic voting that mixes paper and electronic votes. Later I refined it to allow voting over SMS.

In brief, voters are issued paper ballots in the mail or at polling stations in the weeks leading up to an election. These ballots, for each race, show a short code number beside each choice, and a second response code number. You can go up to any terminal, even a hacked, corrupted one, and vote by entering in the code number for your choice. The computer must then show the response code and if you see it, your vote was counted at the central server -- if we can trust the central server. Voters are allowed to request extra ballots and declare only one of their various ballots to be functional. This lets them pretend to sell non-functional ballots to those who would buy or coerce their ballots.

An alternate version of this puts the secrets into a SIM card. You can take the SIM card and put it in any SIM phone -- yours or a friend's or one down at a shop, and run the election app to cast your vote securely, as long as the SIM is not hackable.

This system would work, though it's not perfect. Parties would generate their own voting apps to encourage their faithful to vote straight up and down the ticket, discouraging dissent. People would still hack into computers to tweak the UI to encourage certain candidates. People would do fear campaigns -- as they already do with other systems -- to sway how people vote. Work would need to be done to fight these and other issues.

As noted, this still needs to trust the central voting server, which among other things knows the secret keys behind the ballot codes. Naturally we would work extra hard to secure this machine and its keys, but we would also ask voters to mail in their marked paper ballots as well, though they can mail them after the election. Then, random audits can be conducted to assure that these ballots were correctly recorded. (An additional check number from the server would assure that fake ballots are not considered in this.) More challenging is reliably detecting the creation of false ballots for people who never voted, but this can be mitigated with states pulled during the day.

The result would be quite dramatic. I believe a lot more people would vote. Combined with a fine for not voting and an easy way to cast a null vote, turnout could approach that of the more dedicated countries (like you, Austria with your 92%.)

And GOTV would be a thing of the past, and candidates would need to go back to winning votes by swaying voters, which is how it's supposed to work. Of course they would continue to raise money to try to sway those voters, which is what I'll address in other elements of this series.

Making voting too easy

There is one counter-factor worth considering. If voting is super easy, it becomes easier to convince somebody to vote. If you get easy voting but still low turnout, GOTV becomes cheaper. Of course, if GOTV becomes very cheap, then that possibly is positive because candidates don't need to raise a lot of money to do it, and the real issue is that they feel they must raise lots of money to be elected. The issue with money in politics is that as long as money can be turned into votes, all candidates are forced to spend and raise lots of money. If votes are cheap, and a modest sum can GOTV all the low-hanging fruit lukewarm voters, there is no longer a big difference bought by raising a lot.


I think my all-time favorite alternative voting proposal is to have a "None of the Above" option in every list of candidates. If "None of the Above" wins, either the position is abolished or all the current candidates are disqualified and a new election is held with all new candidates.

On SMS: It seems to me that voting over SMS might make it easier to bribe people to vote a certain way and see who they voted for. You could throw a party where people bring their phone and are pressured to publicly choose a particular candidate slate - without in-person attendance at little curtained booths it would no longer be the case that voting was anonymous by default. It could turn into a group activity - you and your friends have a "voting party" where you all get together over drinks, all run the app on respective phones, discuss, and make your selections simultaneously. Might not be a deal-breaker, but worth thinking about.

Regarding the "blank ballot" option: I stopped voting a decade or so ago because I realized I have moral objections to voting; I actively don't wish to send the symbolic message that I believe voting sends both to other people and to myself. I would like to be able to click a single "I refuse to vote" button and thereby be let off the hook. Sort of a "conscientious objector" status. The right NOT to vote is important to me and if I have that option I am inclined choose it. I would do that even if it were a bit LESS convenient to NOT vote. Your analysis doesn't seem to take that sort of view into account.

Yes, I think you want to support the ability to not vote. However, with the compulsory voting rule, you still would have to overtly declare that somehow -- online or in person.

The SMS system, like my paper system, includes an ability to request "fake" ballot codes. You would get a set of ballot code sheets, and each would have a magic code word or icon at the top. Only you would know which icon signifies the working ballot. Nobody would know how many sheets you got, and in fact everybody could be default get a random number. Again, you do all this long before the election, when you have time to visit the office and get the sheets. (Any office, not just your local one, could give you the sheets, though they would need to coordinate to not issue you two valid ones.)

Invalid codes still appear to work, but the voting system knows to discount them.

Note that vote-by-mail, which almost all states and elections have these days (Oregon is 100% vote by mail) gives up the secret ballot concept already. Unless you do my fake ballot system. So anywhere with vote by mail readily allows vote selling and vote coercion. By adding fake sheets, I actually take that away.

I am not sure where you got your "81%" voter attendance figures for Australia.

The numbers for Federal Elections are typically much higher 93-95%


Institute for Democracy and Electoral Assistance, 1943-2013
Australian Electronal Commission, 2010
Australian Parliament House, 2014

I guess I should not trust the wikipedia folks!

"Are you being pressured to vote in a particular way? Call this number and we'll send you another form that will look identical to this form, but with entirely worthless codes. Only you will know which set of codes are the real ones."

Another scenario to consider is the "dominant husband". The voter is in a relationship with a MITM. All communication between the vote-counter and the voter goes through that MITM who will know which set of codes is the real one. Under old-school voting, the voter at least has an opportunity to directly communicate with the vote-counter.

(I imagine the voter in this scenario has bigger concerns than their ability to vote.)

That would mean complete control of the voter's life. That the voter can't pop into an elections office at any time during the election cycle to request new ballots without being monitored by the dominator. So it's a high bar. Note that all races that allow vote-by-mail allow this sort of domination or vote buying at present. Not that this is a good thing about them, just a sign that vote-at-home is already here.

Compulsory voting with penalties is likely to increase the voter turnout statistic, granted. But is that really a statistic that needs to be fixed? It sounds like you're trying to improve the symptom instead of the disease. People who don't vote have their own reasons, whether apathy, skepticism, or not finding a significant difference between the likely winners of the election. You can force more citizens to vote in an election but you can't make them care more. Voters who take the time to inform themselves are much more likely to vote, especially if the election is likely to be a close one, or is very important to them. Conversely voters who aren't particularly interested in learning the facts are more likely to be swayed by the ads and slanted media coverage (Problem 2). In engineering terms the former could be termed "signal" and the latter, "noise". In many elections, partisans try to "amplify" the noise favoring their side through GOTV (Problem 3).

Part of the reason it pays for organizations to spend vast amounts of money (Problem 1) in politics is the leverage to be gained through government spending in their favor. Whether it's bank bailouts, legislation to bail out union pensions, or defense contracts, having a politician in one's pocket is a significant asset. In business the strategy still holds: a computer software company which sells to banks will try to influence the decision-makers in these banks, just as lobbyists try to influence politicians. The long term solution to "money in politics" is to reduce the payoff for this influence by reducing the amount of money thrown around by the government. That's why there's much less money involved in election of dog-catchers than there is for US members of Congress.

Technical/political Objection: Already there are objections over requiring photo ID for voting, even if provided gratis. Dividing the electorate into those with and without cellphones would likely meet up with the same objections. A closely matched election, such as the 2004 US Presidential race would amplify the claims of discrimination.

As noted, a SIM/SMS style voting option allows you to use any phone or computer. You need not own one, you just have to know somebody who does. And you can still stop by a polling place to use their device of course. With a penalty, people would find a way to find somebody with a cell phone.

Another option would be to require carriers to offer free service for voting, as they provide free 911 service. Ie. you could use a non-activated phone -- it is a new SIM after all. Old phones without activation are thrown out as garbage in huge volumes and could be made available free -- they are already quite cheap.

The non-voters include those who don't even bother to learn, but also lots of people who do have an opinion but don't vote for a wide variety of reasons, including not just the difficulties of voting, but a belief that their one vote will not change anything so why take the effort?

I do agree that if would be better if we could fix the fact that money buys power because government has power. That's orthogonal, though, and frankly, pretty challenging to make happen outside of Libertaria. So the most effective step at present is to make it so that politicians don't need to sell themselves to get elected. They might still sell themselves, of course, but less legally, and because they want to, rather than they need to.

I agree; this addresses the symptom and not the cause.

Many people don't vote because they believe that the real decisions are made elsewhere and/or that they don't really have a choice, which is probably the case in some countries. The main reason for this is the two-party (or, in dictatorships, the one-party) system. Get PR and things will improve. Don't know what PR is, gentle reader? That's part of the problem. Most US citizens probably don't even know what it is. That shows how far the USA has to go in order to become a reasonably good democracy.

I do talk in my New Democracy section about other styles of legislature including more proportional representation.

However, for the specific problem of the world's most expensive democracy, solutions must kludge a bit within the constitutional limitations.

Yes, voters don't feel they can make much difference. This feeling may be correct. But this feeling is at the core of GOTV. The GOTV workers call you up and say, "Hey, we really need you to vote today" and it makes some of them vote. This allows candidates to turn money into votes in the most cost effective way, and that means corruption.

The voter motivation problem is in my view a lack of expressiveness in the ballot. Ranking of choices, a wider range of people elected, brings out votes that otherwise stay home. Without necessarily electing a lot of fringe parties, but at least, understanding the support and appeal and who else they accept to lead.

A generation of experience with alternative vote (AV / instant runoff) or single transferable vote (STV) may be required before we really see whether the lack of expression and choice is what demotivates the voters in any given place.

Party proportionality is not the only goal of a voting system. There is a reasonable list of values that voters in the British Commonwealth expect from a voting system that has resulted from study and numerous electoral referenda in Canada:
lists seven criteria and sub-criteria (interpretations of the main criteria) as follows:

1) broad proportionality,
1a) instant adjustment to ensure proportionality so every vote counts exactly the same without districts having to change

2) extended voter choice,
2a) a fair chance for community-based independents - including municipal politicians - without strong (or any) party ties

3) stable and responsive government, and
3a) aligning municipal divisions more closely to the ecosystems they manage and to stable federal/provincial districts

4) maintaining a link between representatives and geographic constituencies.
4a)guaranteed fixed and stable divisions of political responsibility for ecosystems protected and used in common

5) a relatively predictable election schedule so that small parties or those choosing leaders are not unduly disadvantaged
5a) fixed election dates guaranteed by statute that can only be over-ridden by a specific protocol that delays the writ drop.

6) simplicity insofar as an ordinary member of the public can learn and practice the counting system with no expertise
6a) zero tolerance for any system that does not guarantee a paper trail, paper recounts, or relies on electronic storage
6b)backwards compatibility of the ballot so that those familiar withi FPTP voting systems can participate without spoilage

7) the public themselves, not the political parties, makes the final choice regarding who is to be trusted with real power
7a) every individual who acquires a vote in the legislature must have been selected by a process open to public input

[the article has more rationale for each point and interpretation]
shows some options for different kinds of ballots explains the superiority of the vote swapping method for small party supporters, and why it is not related at all to the "strategic voting" advised by the mass media
is a description of a voting system that combines STV and MMP attributes without some of the arbitrary thresholds that MMP is prone to (and which were, at least in Canada, overturned as arbitrary and thus unconstitutional in the context of funding parties) and without the objectionable "party list" component, and which satisfies the "voting system values" as above.

It is common that American voters can't find candidates who motivate them to vote. The two-party system looks like the bigger problem. There are presently zero independents in the House. Zero.

Geographic districts already have representation in the state houses. What good is district level representation in Congress? It certainly prevents new parties from developing, tends to stifle innovative ideas, and encourages pork barrel spending at the federal level.

I agree,but "independent" as an alternative to the two established parties is just a step in the right direction which is OK as far as it goes, but doesn't go far enough. One needs not just Democrat, Republican, and Independent, but Democrat, Republican, Green, Socialist, Communist, Liberal, Libertarian---whatever people want to vote for.

And proportional representation in Congress would enable reps from smaller parties to get elected.

I find the idea appalling. Just what we need cancelling our votes - people who can't be troubled to turn out otherwise.

I'd go the opposite direction: make voting a serious duty that you don't undertake lightly. Do away with any voting other than at your neighborhood polling place, except with very good excuses. Let's turn voting back into a civic social event, rather than a sterile button push from wherever you happen to be. And for heaven's sake, let's not encourage people to vote.

Voting only by those who are dedicated and serious is simply not an option in today's system. And I don't think your proposal would change that. This is the point I am trying to make. You might like that as an option but it's not available. The options I see are:

  1. People who care a bit and whose lives and schedules make it practical to get to polling places vote. In addition, a deciding volume of people who would not have voted are converted into active voters through GOTV efforts and money.
  2. Voting is made easier, and those who care vote, and those who wouldn't have voted if it weren't easy also vote. GOTV becomes less productive if turnout is high, or becomes super-productive (and thus not that expensive) if turnout remains low without it.

I am not certain if #2 is good but #1 seems bad, because it magnifies the ability of money to buy elections.

Attempts to do what you suggest, make voting hard, tend to usually be associated with voter suppression. People who have long hours or two jobs, couples where both parents work, people without cars, or who have many kids are less likely to vote, no matter where the difficulty bar is. It's not that they don't care, it's just that if there is a bar, any bar, then some people will get over it and some will not. And those who do not are prime for GOTV efforts. Even if you make the bar really high, the only question is how many people GOTV can transform from just below the bar (not voting) to just above it (voting.)

If everybody votes, there is no GOTV. If the only people who don't vote are those who truly have no desire to vote, then GOTV is not very productive.

No amount of technological innovation can fix non-paper ballots. Here's why:

1. Scrutineers. Every party has a representative that scrutinizes the polling process and counts and how any given voter is qualified/disqualified. These cannot be distributed to all physical locations involved in any electronic vote, so online voting means giving up the scrutineer role entirely. Very dangerous and against the explicit laws of most countries.

2. Polling place neutrality. A physical polling place is far easier to patrol than millions of browsers or phones. It's an important principle of fairness that no last-moment intervention and especially not a lie can sway a vote. With a few closely watched physical places with scrutineers of all parties, this is at least possible to begin to enforce even in new democracies. However, nothing can possibly reliably detect a last-moment popup from a candidate misleading the voter on their own phone or browser. An electronic polling place is a non-neutral polling place, falling to whoever has more technical facility with viruses, popups, worms and so on.

3. Recounts. The means by which anything but a paper ballot is recounted cannot be explained reliably to a majority of the public. No amount of mathematical proof will convince a doubting supporter of the losing side that the recount was real.

No matter how complex the voting system, an ordinary person with primary school education can be taught how to count paper ballots in it within an hour or so. The system need not be simple at first glance but it *must* be apprehended by the primary school educated person. Nothing based on software, operating systems, mathematical proofs, ever could be.

Any electronic election is thus an election in doubt.

4. Physicality. It's a more abstract concern but a vote is how we collectively direct violence: a force monopoly held by the state. People will be arrested, jailed and imprisoned, and others will be freed to do things we may object to, based on how we vote.

Why should this kind of decision be made anywhere but in public, exposed to the physical reality of the neighbors and the place we live? Why should it be like shopping, a "personal choice", made behind safe guarded walls? The people most vulnerable in society to our voting choices have no walls at all.

If you cannot brave the street and your neighbours to go vote, perhaps you really should leave voting up to your neighbours who do in fact go out in the street.

Disabled, shut-in, ill, elderly people might well be exempted from this physical movement requirement... but for those of us who *can* get up and go, there's no excuse for not taking one good hard look at our neighbours and neighbourhood before deciding how the violent force monopoly ought to be directed.

5. Insufficient testing. Perhaps the day will come when the ordinary person feels they do understand the mathematical proofs and is ready to discard all the checks and balances of scrutineers and polling places, but even if so... the first place to test a totally electronic voting system is *within parties* not in a general election.

Nothing prevents a political party from selecting its leaders or policies based on whatever innovative voting scheme it wants. Let's see if what comes out of that process is acceptable to the public or not, before we commit the public to that process without any kind of test.

Give it 20 years before we discuss electronic voting again... plenty of time to try it in all sorts of less dangerous contexts than a general election for representatives of a developed world democracy.

Most of those are true. But it is more complex.

Digital voting would have different types of scrutineers. You still would watch the people who write the software and install it. You would watch the results coming in and audit them. There is no need to audit how well the counters are checking for spoiled ballots and ruling on them because there are no spoiled ballots, or if there somehow are, it's not a judgement call.

Polling place neutrality. In Canada, it's fine. In other places there are goons with machetes near the poll trying to intimidate you. And for some people it's a one minute walk, for others much further. In the USA some places have 2 hour lines and many people give up, or polls have to stay open after 8, and people don't get time off to vote in that country.

I agree that the recount concept is difficult with electronic voting for the reason you say.

The system I have proposed retains some physicality, and audit capabilities. To do a full recount all would need to mail in after, which they would not do, so it's a challenge.

As for testing -- actually almost all other elections in the world done by distributed groups are done in electronic form now. Elections of boards of directors, society officers etc. Of course, these are not high value targets in most cases so they are done quite insecurely. But it's very common.

> “Elections should be decided only by those who care enough to make the effort” just isn’t an option on the table.

I'm not necessarily advocating this, but one way to do this would be to make voting *harder*, not easier. Make the forms onerous enough, and GOTV becomes ineffective because mild supporters won't make the effort.

There is a reason you say you are "not necessarily advocating this" as the history of making voting harder has unfortunately been a history of attempts at voter disenfranchisement.

Generally we are pretty clear on the idea that poll taxes, racist voter ID laws are attempts at disenfranchisement. It is much less common for us to see whatever stops 40% of the electorate from voting as disenfranchisement. We generally just view it as apathy. But it's often a case of whatever barriers there are (including learning about the race, and having time to get to the polls) winning or losing against desire to vote. The fact that any one vote has a ridiculously low chance of affecting the result certainly adds to this.

I wonder if having multi-candidate voting (like Approval voting) might bring more people out. They will know they can't change the results of the big race, but the chance to bump the percentage for the minority ideas they support could be a big deal.

Add new comment