The turnout narrative

Over the past few weeks, I have heard a specific narrative about turnout many, many times.  Most people argue that high turnout is good for the KMT because DPP supporters always turn out to vote while KMT voters are more fickle.  Sometimes they will vote, and sometimes they won’t.

I suspect that this narrative is completely wrong.

 

Turnout is an incredibly hard phenomenon to study.  We all know that mobilizing your voters is a key part of the election game.  If you can get a high percentage to the polls and your opponent can’t get many of his supporters to vote, you have a huge advantage.  However, the final number that we see is the combination of both these groups mixed together.  You can’t really study this with opinion polls either.  For whatever reason, the results you see in opinion polls don’t look anything like the actual election results when it comes to turnout.  Reported turnout is always much higher than actual turnout in both pre- and post-election surveys.

That is a way of saying that I don’t have any good evidence to back up my ideas, but neither does anyone else.  I think the lack of hard evidence is part of what is driving the narrative.

 

I think the turnout narrative is driven by lazy thinking.  It emanates primarily from people in the blue camp (though I have heard it repeated by people in the green camp as well), and I think it reflects a blue camp perspective of the turnout problem.  People always have a tendency to think that their own problems are more daunting than those of their opponents.  For blue politicians, turnout is one of the major components of a traditional campaign.  You put together a large, complicated, and expensive grassroots network of people, and you spend a lot of energy and money motivating those people to go to the polls and vote for you.  Blue politicians must look at the green camp with envy.  Green politicians traditionally have much, much less organization at the grassroots level, and yet their voters somehow show up at the polls.  It would be easy to conclude from this that green voters are ideologically motivated (read: fundamentalists) and always show up to vote.  Blue voters, in contrast, require a lot of work.

This narrative has blue supporters wringing their hands in anxiety.  As one radio commentator put it, “If the weather is bad, people will stay home.  If it is too good, they will go out and play somewhere.  Why can’t our voters be more like those fundamentalist DPP voters?”  Note that this also makes a nice message for the last few days of rallies: the other side is going to turn out in force; we have to match their intensity.

 

Several years ago, I seem to remember a similar message, but it came from the other side.  It was the KMT that had iron voting brigades, and DPP supporters wondered how they could ever overcome that obstacle.  In fact, the KMT military or government villages made up only a small part of the electorate in most districts, just as the DPP fundamentalists account for only a portion of the electorate today.  Everyone tends to emphasize the difficulty of their own challenge.

I recently spent a few hours combing through Taipei City electoral data looking for evidence of a rock-solid DPP vote and a variable KMT vote.  I looked at data at the li level from 2002, 2004, and 2006.  This includes high and low turnout elections and good and bad KMT elections.  If this turnout narrative is correct, there should have been some discernable patterns, with strong KMT li behaving differently than strong DPP li.  Now, as I said before, turnout is notoriously difficult to analyze, and I doubt I got everything specified correctly, so I’m not going to present any data here.  I will just note that every time I managed to find a statistically significant pattern, it was in the opposite direction as expected.  In other words, every time something showed a clear pattern, it seemed to indicate that higher turnout helped the DPP, not the KMT.  However, I’m not at all confident that I did any of this correctly, so let’s leave it at that.

A quick glance at the data is probably better evidence against the narrative.  The total number of blue camp votes goes up and down in various elections, but so does the total number of green camp votes.  I simply don’t see any rock-solid green vote that always turns out.  Some of the highest turnouts in Taiwanese electoral history produced pretty good results in Taipei for the DPP, including the 1998 Taipei mayoral campaign and the 2004 presidential campaign.  The 2002 and 2006 mayoral campaigns, in which the DPP did relatively poorly, had relatively lower turnout.

In fact, the DPP spends a lot of energy on turnout; they just do it in a different way from the KMT.  Instead of building a grassroots organization that will contact individual voters, the DPP excels at wholesale politics.  All those enormous rallies are not about changing minds, they are about increasing turnout.

At any rate, I believe the general proposition that high turnout is good for the KMT because KMT voters are more fickle is a lousy one.

 

That said, I don’t necessarily believe that high turnout is not good for the KMT THIS YEAR.  This year, I suspect we are looking at a DPP wave, with energized DPP supporters and lackluster KMT supporters.  This year, enthused DPP voters will probably turn out en masse, and the overall turnout rate will probably depend on how successful the KMT is in turning out its more begrudging supporters.  A moderate turnout of 65% would probably spell disaster for the blue camp.

In other words, this year looks to me a lot like 2008 in reverse.

 

Reforming the voting process

Taiwan is considering changing its rules on how to vote, and this essay will consider the possible benefits and costs of some possible changes.  The changes being bandied about are not wholesale changes to the electoral system, but rather changes in how, when, and where voters can cast a vote.

In a nutshell, the problem is that Taiwan’s current system is quite inconvenient.  Voters are automatically registered at their place of official residence.  On election day, they must vote in that polling station.  There is no provision for people to vote early, by mail, or in another place.  In other words, if you are away from home, you have to return home to vote.  Large numbers of people do not live in their official place of residence.  This includes groups such as students, active military, people living overseas, and people who simply work in one place while maintaining their official residence in another.  Every election day, massive numbers of people jam the highways and trains to get home to vote.  You also cannot vote if you cannot actually get to the polling station.  People confined to sickbeds, in prison, active duty military personal, or whose jobs require them to be on duty all day are prevented from voting.  It is a very simple and restrictive system, wonderful for bureaucrats and sometimes onerous to voters.

Of course, this has been the system for several decades, and Taiwan’s voters have still managed to vote at reasonably high rates.  During the 1990s, turnout was often in the 70-80% range, peaking with 82.7% turnout in the 2000 presidential election.  However, turnout has dropped in recent years, and high 50s or 60s are not uncommon.  One reason for this is the maturing of democracy.  Democratic elections are simply not novel any longer.  In the 1990s, every election seemed to be a historical first: the first full election of the national assembly (1991), the first full election of the legislature (1992), the first direct election for an island-wide position (1994), the first direct presidential election (1996), and the first real possibility that the opposition would win power (2000).  There is probably also some disillusionment with politics after the scandals of Chen’s second term.

However, the real force driving this debate is that a large number of Taiwanese now live in China.  I have heard a wide range of estimates, ranging from a few hundred thousand in all of China to two million in Shanghai alone.  Since Taiwan’s population is just under 24 million, the latter figure would mean that 8% of Taiwan’s population resides in Shanghai, with perhaps double that figure in all of China.  Of course, since the Taiwanese in China are presumably disproportionately adults, the percentage of the electorate would be even higher.  This doesn’t seem realistic to me, but no one knows for sure what the actual number is.  The only thing we are certain of is that the number is very large.

Not only is the number of Taiwanese in China large, but everyone assumes that these voters would overwhelmingly vote for the blue camp.  Businessmen are subject to myriad pressures from China.  And you wouldn’t go to live in China in the first place if you weren’t sympathetic to the idea that you are Chinese.  Most rabid Taiwan nationalists do not consider China a suitable place to go seek their fortunes.  Or so the conventional wisdom goes.

So the problem is complicated.  There is clearly a democratic deficit.  Lots of people are effectively disenfranchised.  Whether this is 5%, 10% of the electorate, or even more, what is beyond dispute is that there are many people who cannot vote under the current system.  Within Taiwan, there is probably a class distinction, with many lower income people living away from home.  (For tax purposes, you maintain your official residence in a rural area where you own a home, but you rent in the much more expensive Taipei real estate market where you work.)  Outside of Taiwan, there are large numbers of people for whom voting entails a very large cost.  Universal franchise is a basic component of almost all modern definitions of democracy, and Taiwan may be falling short.

However, the solutions may be worse than the problem.  Most solutions are open to abuse of one sort or another.  As a new democracy with some very contentious and controversial elections in its near past, Taiwan’s first task is to hold free and fair elections.  Not only do they have to actually be free and fair, they have to be widely perceived as such by almost the entire citizenry.  This is a very high bar, but it is simply the nature of democratic elections that all losers are sore losers.  If they can plausibly claim they were cheated of their rightful victory, they will almost always do so.  The electoral process has to be made so foolproof that losers simply cannot plausibly complain about anything except the stupid voters who voted against them.  Anything less will cause the democratic system to start unraveling.

So what are the options for Taiwan?  Several ideas are being mooted, including (1) allowing voters to vote in places other than their official residence, (2) early voting, (3) proxy voting, (4) voting by domestic mail, and (5) voting by international mail.  To this, I will add another possibility, (6) setting up voting stations overseas.  Let’s go through these one by one.

1. transferring registration

Currently, the only exception to the rigid system described above is for election workers.  Election workers, who staff the poll stations all day and would thus not able to vote if they lived in a different area, are allowed to transfer their registration to the poll station in which they work.  They have to apply for this sometime previous to Election Day.  So the proposal is to simply expand this system, allowing anyone to transfer their registration.

This is the simplest potential change.  While the government has researched a wide variety of changes, it is currently only suggesting implementing this one right now.  The Interior Ministry is suggesting that it might be possible to allow registration transfer in December’s election.

While this is the least controversial of the potential changes, the DPP is objecting.  I believe that the biggest reason for their objection is not really that this is a bad change, but that they are afraid of the slippery slope.  They are really afraid of the other changes, and the best way to forestall those is by stopping this one.  That said, there are legitimate concerns about even this proposal.  The two most serious questions concern where the polling stations would be and whether the freedom to vote without coercion might be infringed.

First, would polling stations be set up inside institutions such as police stations, military bases, or prisons?  Since these institutions already have well-established limits on freedom of action, it would be difficult to guarantee the integrity of voting.  For example, even if opposition[1] party observers were allowed inside, they might experience limitations on their freedom of movement, their ability to watch everything they deem critical, or on their ability to communicate with the outside world.  Moreover, even if everything were actually conducted fairly, a large percentage of the population would be skeptical.  I had originally assumed that this was such an obvious point that no one would seriously propose setting up polling stations inside sensitive areas.  However, the Minister of the Interior said in an interview that, under the new system, President Chen would be able to vote from within his jail cell.  This is exactly what I am worried about.  I do not worry that prison officials would be so stupid as to try to coerce Chen to vote against his preferences, but there are thousands and thousands of other prisoners who are much more vulnerable to pressure.

The obvious solution to this problem is using the normal polling stations outside these special institutions.  So instead of taking two days to go home to Tainan to vote, a soldier in Jinmen could take two hours, go outside the base to the nearest elementary school, and vote at that precinct.  There would be no military police on guard, no commanding officer or political officer ominously loitering around, and far fewer limitations on party observers.  However, even here there is a potential problem.  When soldiers go home, it is impossible to know if 90% of them voted the “right” way.  If all soldiers at a base (or in the same unit) go to the same precinct, it is not hard for the commanding officer to learn how the unit voted collectively.  Army officers are very good at enforcing collective responsibility.

Perhaps the solution is to assign soldiers, police officers, and members of other groups in which there is the possibility of collective responsibility randomly to precincts.  In other words, the soldiers in a particular unit would be assigned randomly to precincts all over Jinmen, so that no single commanding officer would be able to look at a particular ballot box and figure out how his soldiers had voted.  This would increase the cost of voting for individual soldiers, but it might be necessary to allay fears of manipulation.  Again, the danger is not simply from actual abuses, but also from popular suspicions of abuses.

Case Study: Yuli Mental Hospital

In Yuli Township, Hualian County 花蓮縣玉里鎮, there is a hospital with a large wing for patients with mental disorders.  On the hospital’s website, there is the following (undated) press release about how patients happily and freely exercised their right to vote:

http://www.ttyl.doh.gov.tw/main.php?index=newtxt&pid=&page_name=detail&pageNo=1&iid=9&no=

The gist of this press release is as follows:  In the past, every election saw the appearance of voting brigades of two or three hundred people who only knew the number, not the name, of the candidate they would vote for.  Since 2003, the hospital has reformed its administration to concentrate on our core mission, and we are resolved not to get involved in local politics any more.  We have required that the voting booths inside the hospital be removed so that patients who vote will be able to go into the town and vote as part of the community.  We also require strict neutrality from our staff.  Patients are now freely allowed to decide whether they will vote or not.  As a result, instead of the two or three hundred voters that we had in the past, only 64 patients voted in this election.  All of them were able to temporarily leave the hospital to go into town and vote for the candidate of their choice.

The hospital is located in Taichang Li 太昌里, and past elections certainly look suspicious.  For the record, all of these candidates except Zhang Wei were KMT nominees.  Zhang Wei was a KMT member supported by the Huang Fuxing (military) system who ran against the official KMT nominee.

year	type	candidate		Box #1	Box #2	Box #3	Box #4
1995	Legislative	Zhang Wei	張偉	38%	73%	76%	3%
1995	Legislative	Zhong Lide	鍾利德	26	6	4	78
1998	Town mayor	Pan Fumin	潘富民	32	78	88
1998	County Ass.	Yan Tingxiang	閻廷相	38	7	81
1998	County Ass.	Chen Rongcong	陳榮聰	6	68	0
1998	Legislative	Zhong Lide	鍾利德	34	4	57
1998	Legislative	Zhang Fuxing	張福興	28	84	37
2000	President	Lian Chan	連戰	23	84	83
2002	County Ass.	Yan Tingxiang	閻廷相	43	2	67
2002	County Ass.	Liu Dezhen	劉德貞	24	88	14

The first box is almost certainly the normal residents of the areas around the hospital.  They look a lot like the results from the rest of Yuli Township.  However, the other boxes are fairly extreme.  KMT candidates get far, far higher percentages of the vote than in surrounding areas.  We even find the KMT splitting the votes between two of its candidates, a sure sign of a well-organized machine.  After 2003, the votes no longer look like this.  Instead, all the boxes look like the first box.  It certainly looks as if the administrative reform referred to in the press release was not merely lip service.

This example illustrates my argument quite well.  Setting up voting booths inside an institution with coercive potential is open to abuse.  Voters are right to be wary of these abuses because they clearly have happened.  Two things were at the heart of the hospital’s reform: they ordered their staff to be neutral and they moved the voting booths outside the hospital and into the wider community.  The importance of the second point cannot be overstated.

There is also another point, one about perceptions.  I first heard of this hospital and its voting brigades nearly two decades ago when I was visiting a friend in Hualian.  Her father was convinced that the DPP could never win an election there because of this hospital.  He believed the elections were so stacked in favor of the KMT that the DPP could never overcome this disadvantage.  When I went home and looked up the numbers, I was flabbergasted to find that it was only a few hundred votes.  Now, a few hundred votes is not nothing, but the DPP has never come within a few hundred votes in Hualian.  Likewise, a few hundred votes would rarely prevent them from winning in places where they are popular.  The point here is that the perception of unfairness was much worse than the actual reality.  This is the normal way of things, and this is precisely why it is so critical for the electoral process to be beyond reproach.  Voters do not look at abuses and think that the election was 99% fair.  1% unfair is blown into 100% unfair.  The government has to be very careful with its proposed reform so that in addition to making it nearly impossible for the Yuli Hospital to influence votes, it is also very difficult for reasonable people to argue that the hospital could influence votes.

2. Early voting

Early voting is not the same as American-style absentee voting (which will be covered in the voting by mail section).  Early voting refers to a system in which the voter can go to the normal voting booth and cast his or her vote during the period before election day.  So instead of a single day, the polls would effectively be open for a week or ten days.  The problems with early voting are mostly logistical.  If the Central Electoral Commission can figure them out, then early voting should be fairly uncontroversial.

One problem is getting the election gazette and ballot papers printed in time.  As it currently stands, printing cannot start until candidates draw numbers two weeks before election day.  The gazettes are distributed a few days before the election via the large network of village heads, and ballot papers are usually shipped out to each precinct a day or two before the election under heavy security.  This schedule would have to be condensed.  Perhaps candidates would simply draw numbers a week earlier and everything could be moved up a week.

Another problem involves the voting stations themselves.  Would the government really set up all ten thousand voting stations for an entire week?  This would be nearly impossible, especially since most voting is done in schools, and the schools are occupied during the week.  Also, a high proportion of election workers are schoolteachers, and they are also busy during the week.  What would be more likely is that there would be one booth per area (perhaps one per township) set up in a local government building.  But then voting rolls would have to be rigorously checked to prevent double voting.  A single precinct has one single list of people who can vote at that precinct.  With overlapping lists, the potential for confusion and abuse grows.  Again, this is a logistical problem and simply requires a careful solution.

Ballot security becomes more difficult with early voting.  When voting occurs in a single day, election observers can see that the box is empty at the beginning of the day, and they can continuously watch it all day to make sure that no one puts extra ballots in.  (In some countries, plexiglass boxes are used to make stuffing the ballot boxes even harder.)  Then they are counted before anyone leaves, so observers can be fairly confident that no one has tampered with anything.  With early voting, ballots would have to be stored overnight.  The system would have to be sufficiently secure for both sides to be confident that there was no way the other side could tamper with the ballot boxes during the night.

There is one point that is not just logistical.  Is there a value in having a common campaign period and a single election day for all voters?  Recent American campaigns have had to adjust their schedules to account for the fact that a very high proportion of people vote early.  Taiwanese campaigns typically hold huge events on the weekend before the election and on the night before the election.  In 2000, the Chen campaign purposely timed the Li Yuan-tse endorsement so that it would have just enough time to sink in before election day.  I’m not sure this should be a major consideration, but early voting would change some of these schedules.  Also, as an elections nut, I quite like the excitement and bustle of a single, common election day for everyone.  My inner anthropologist tells me that common rituals are critical to building communities.  But I suppose that wider participation in these common rituals is also quite important.   In all, I think early voting is probably a reasonable idea if it is carefully implemented.

3. Proxy Voting

Proxy voting refers to a system in which one voters gives another person the power to cast his or her vote.  For example, if a person will be out of the country, she can empower her husband to cast her vote for her.  In general, proxy voting is a bad, bad idea. It is by far the worst of any of these potential reforms.

Logistically, I’m not quite sure how proxy voting would proceed.  In the lax version, a person could empower a proxy simply by handing over his or her ID card.  Anyone in possession of another person’s ID card would be legally empowered to vote on their behalf.  In this case, I imagine widespread theft or counterfeiting of ID cards.  In a stricter version, a voter would have to apply in advance to empower a proxy by signing some paperwork at a government office.  Either way, the potential for abuse is legion.

Proxy voting would instantly make machine politics much more practical, effective, and coercive.  The biggest problem machines have is that vote buyers cannot verify the actual vote to ensure that the voter has followed through with his end of the bargain.  With proxy voting, a voter would be “encouraged” to sign over proxy power to a trusted accomplice.  One can imagine a precinct in which 20 village heads each show up with 50 ID cards in hand and the result is 1000 to 0.  A voter who did not sign over proxy power would immediately be singled out, and the appropriate pressures could be applied.  Good old-fashioned clientelism, vote buying, and simple intimidation would rule the day.

Proxies could be limited to immediate family members.  This would mitigate but certainly not resolve the problems of machine politics.  Instead, the problems would be recreated within the family.  One might imagine the family head insisting that all other members turn over their right to vote.  Since the most powerful person in most Taiwanese families is male, this would effectively disempower large numbers of female voters.  It might also disempower voting-age children, especially those not living with their parents or those who are not financially independent.  On the flip side, seniors dependent on their children might also lose their right to vote.  Candidates would immediately react by downplaying issues that women, 20-somethings, and seniors care about.  Again, proxy voting is just a bad idea.

4. Voting by domestic mail

Voting by domestic mail only is probably not a serious proposal.  If they were actually to limit voting to only domestic voting, they could accomplish nearly everything through registration transfers and early voting.  Also, it would be strange to allow only ballots arriving in envelopes with domestic postmarks but invalidate those arriving with international postmarks.  However, I wish to delve into problems involving mail without thinking about international mail, so this is a convenient way to sort through those problems.

Americans are familiar with voting by mail, so they tend to assume that it is problem-free.  Taiwanese see that Americans do it, so they ask why they can’t.  I, on the other hand, am continually amazed by Americans’ willingness to overlook all the potential abuses that come with voting by mail.

First and foremost, there is the problem of ballot security.  How do you know that the ballot that is mailed out to a voter is actually received by that voter?   This could easily turn into proxy voting.  Imagine a large household with several votes, all of which apply for mail ballots.  The person who collects the mail has access to all these ballots.  In the American system, the voter supposedly has to sign the ballot to affirm that he or she has personally marked the ballot.  Signatures can, of course, be forged since no one is going to actually check each and every one against some central database.  In Taiwan, this would be replaced by a personal seal.  Since anyone with access to the seal can use it, this would do very little to ensure against proxy voting.  This all assumes that the intended household actually receives the ballot.  Mail routinely gets lost or delivered to the wrong address.  Mail can be stolen.  The postal system could be corrupted.  Currently, the government goes to great lengths to ensure the security of ballots.  They are printed and delivered under heavy security, and each ballot is rigorously accounted for.  Once you allow voting by mail, all that security is immediately thrown out the window.  In a polity in which both sides think the other will cheat if given any chance, is it a good idea to loosen all controls on ballot security?

Another problem with voting by mail is that it requires as much or more time than early voting.  After numbers are drawn, ballots would have to be printed, mailed out, and then mailed back in.  If you allow two days for mailing in each direction and three days for the voter to get the ballot, mark it, and put it back in the mail, you have to have everything ready a week before the election.  Three days for the voters is not very long.  By comparison, American absentee ballots are typically sent out a month or so before the election.  Can you imagine what would chaos would have erupted if several thousand mail votes had showed up on the day after the election in the 2004 presidential election?  That scenario is almost inevitable with voting by mail.  Or perhaps you would allow for all ballots postmarked before election day to be counted, no matter when they were received.  In that case, you might not be able to proclaim a winner on election night.  Since providing a timely resolution to political conflict is one of the prime contributions of elections to governance, this would also not be desirable.

As I said above, there really is no reason for voting by mail if one is only concerned about domestic cases.  Almost all the same benefits could be reaped through registration transfers and/or early voting without the same risks to ballot security.  Of course, it would also be impossible to ensure that ballots were not mailed to friends and relatives in Taiwan from voters overseas and then mailed domestically.  And that would bring us to the real question, that of voting by international mail.

5. Voting by international mail

Voting by international mail would exacerbate the problems of voting by domestic mail, especially in the case of voting from China.

Voting by international mail would take much longer than voting by domestic mail.  Mail to many parts of the world takes over a week each way.  Ballots would probably have to be mailed out a month before the election.  In the current system, this is when the candidates are finalizing their registration, so the entire calendar would have to be changed.  It is possible that even a month might not be enough time for voters in some remote locales, but I suppose we won’t worry about them.

International mail is also less reliable than domestic mail.  The chances of ballots getting lost or delivered to the wrong address are probably much higher.

However, the real problem is that of ballot security.  Even if we assume the absolute integrity of the domestic postal system (and it’s not clear that we should), all bets are off when it comes to foreign countries, especially China.  I am not terribly worried that the government of Malaysia or Japan might try to corrupt a Taiwanese election.  Even if they did, there aren’t that many voters in those places.  However, China is different.  There are a lot of voters, and it would be a shock if the PRC government did NOT try to intervene.  China has strong preferences over the candidates.  It has no reason to stay out, since its official ideology is anti-democratic and it does not consider Taiwanese elections to be legitimate or outside its rightful sphere of control.  China also has the wherewithal to get involved.  Its security apparatus could easily corrupt the mail system, allowing ballots to reach only voters that it trusted.  It could open all the votes and “lose” the ones that were marked “incorrectly.”

This is not merely a problem of making sure that all the votes that are actually cast are in fact counted.  Allowing votes to go through China’s mail system would undermine the freedom of choice.  Because China can open mail, it could easily match votes to specific people and sanction those who chose the wrong candidate.  It would be very easy for the Chinese government to put irresistible pressure on voters to choose its favored candidate instead of theirs.  In fact, the government might not even have to do anything.  Simply because Taiwanese in China know that the government could open the mail and monitor their votes if it wanted to, they might all vote blue just to be safe.

In fact, voting by mail might actually make things worse for Taiwanese living in China.  Currently, at least they can come home and vote freely.  With voting by mail, one can imagine the Chinese government putting pressure on voters not to return home to vote, but to stay in China and vote through the (easily monitored) mail.  Anyone who returned home to vote could be considered disloyal (to China) and sanctioned accordingly.  Simply by having the system of voting by international mail available, the Taiwanese government might effectively strip all Taiwanese living in China of their right to vote freely.

The point about perceptions must also be reiterated.  Suppose China did not actually interfere.  Many green supporters would simply not believe that it had not.  The simple fact of China being involved in any way, even if only through the postal system, would be sufficient to cast doubt on the integrity of the election.  Every time the blue camp won an election, there would be doubt.  And if the green camp won, they might consider themselves to have an overwhelming popular mandate, since it would have had to be big enough to overcome the illegitimate blue votes coming from China.  Overestimating one’s popularity is often a cause for democratic breakdowns.

In short, voting by mail from countries other than China would be problematic; voting by mail from China would be a disaster.  The risks to the democratic system are simply not worth a slightly higher turnout.

6. Setting up voting stations overseas

This is the Mexican solution.  In the 2006 election, Mexico set up poll stations in several locales outside its borders, most prominently in several American cities.  For example, Mexicans in Chicago were able to vote at an official polling station in Chicago.

This system alleviates many of my concerns about ballot security.  The electoral commission could still transport the ballot papers under lock and key to a specific location, such as the Taiwan representative office in Kuala Lumpur where a poll station would be set up.  (I suggest the representative office since any use of a private property might open the door for procedural irregularities or clientelistic pressures.)  The various parties could still have official observers at each station, and voting and counting could be conducted in a single day.  Everything would be overseen by the Central Election Commission, exactly as it is done in Taiwan.  Voters would have to show up in person.  It might be difficult for a voter in Fresno (which presumably would not have a poll station) to drive to Los Angeles (which presumably would), but this would be far less onerous than returning to Taiwan.

The problem, again, is China.  I don’t think the Philippines or Indonesia would object to Taiwan allowing voting in their territory.  However, China almost certainly would.  China does not and cannot admit the legitimacy of the Taiwanese government.  It would be unthinkable for them to allow the ROC Central Election Commission to conduct activities on PRC territory.   The PRC will not even meet with anything but “unofficial” bodies such as the Straits Exchange Foundation, and the CEC is clearly an official organ of the ROC government.  Taiwanese elections are intimately related to questions of PRC sovereignty, and the PRC simply could not stand by idly as Taiwanese decided these questions in an official capacity on its soil without its input.

Moreover, Taiwan does not have an unofficial representative office in China, so there is no obvious place to set up a voting station.

In sum, setting up overseas voting stations is a reasonable solution for the rest of the world.  It would probably satisfy most Taiwanese concerns about fairness in China, but the Chinese government will almost certainly not allow it.

Conclusions

There is clearly a problem here, with large numbers of citizens facing significant obstacles to voting.  However, some solutions are worse than others.  Proxy voting is a terrible idea.  Voting by mail is also not a good idea.  Taiwan should not assume that whatever the USA does is the correct solution.  There is no equivalent of China for the USA.  5% or 10% of the American population does not live in Mexico.  Mexico does not claim sovereignty over the USA.  Mexico does not have a clear preference for the Democrats over the Republicans that is vital to its national integrity.  Mexico might not even have a security system with the organizational capacity or technical expertise to corrupt a mail-in system.  Even if all these conditions were met, Mexico would not dare interfere with an American election because if they did and were discovered, the backlash from the USA would be disastrous to them.  However, all of these conditions do hold for China and Taiwan.  American-style absentee voting by mail is simply not a good idea for Taiwan.

There are some steps that would be good ideas.  Early voting and overseas voting stations should be feasible.  Transferring registration is also a move in the right direction, provided it is done carefully.  Voting booths must not be established in any places in which voters might be coerced, such as military bases or police stations.  The utmost care must be taken to ensure that members of easily sanctioned groups, especially military or police units, do not all vote at the same ballot box and cannot be held collectively responsible for their votes.  With rigorous controls, voting can be made more accessible to larger numbers of citizens.  At every step, the Electoral Commission should considering not only whether the proposal would actually be fair, but also whether it could reasonably be construed as unfair.  Perception are as, perhaps more, important than reality.

Early voting, registration transfers, and overseas voting stations help voters in Taiwan and in other countries, but what about all the voters in China?  After all, aren’t they the ones driving this whole debate?  There just isn’t much of a solution for them.  Early voting helps, since they will have a little more flexibility in when they return to Taiwan to vote.  However, the risks associated with using the mail system are simply too great, and China would never allow ROC poll stations on its territory.  At present, these voters are not completely disenfranchised.  They do still have the possibility of returning home to vote, and many exercise this right.  A system that China could penetrate would actually make them worse off, as China would be able to demand that they vote on Chinese soil according to Chinese preferences.  This also would undermine the Taiwanese democratic system.  So unless China allows Taiwan to administer neutral elections on PRC soil, I can’t see any way to dramatically improve the current situation.  Maintaining the integrity of the process has to be the first priority.

---

[1] I assume that the governing party would be the one trying to rig the votes, since it is the party with immediate control over budgets, promotions, and so on.  I could also assume that the KMT would be the party trying to rig the elections, since most government officials who have partisan leanings are associated with the KMT, especially in the coercive branches of government.  Regardless, the abstract argument is the same.