We have now entered the blackout period, so no new polls can be announced. I want to discuss two questions about the blackout period. First, is it reasonable to have a blackout period? Second, who does the blackout period benefit this year?
Is the blackout period legal? Some people would argue that it infringes on the constitutional right to free speech. I’m fairly certain that the US Supreme Court would strike down any blackout period if the US tried to adopt one. (I’m citing the US case because, compared to other countries, the US courts have historically placed very high importance on free speech relative to other priorities. In other words, the US is the extreme case.) My own opinion is that the state has a compelling interest in maintaining the health of democracy, and that gives it the right to regulate elections. Free speech is not an absolute right; even the US Supreme Court allows free speech to be infringed in numerous contexts. I believe that society’s interest in democracy is sufficiently important to justify some mild infringements of free speech.
How does the blackout help democracy? The usual argument is that when voters have too accurate information about the election, they will not bother to vote. After all, if you are sure who will win, why should you vote? A short blackout period allows for a bit of uncertainty. Conditions might change a bit in the last ten days so that voters are a little more unsure about who will win. Theoretically, this should increase turnout. I think this argument is certainly sufficiently plausible to satisfy any legal requirements.
Fine, that takes care of the legality, but why do practical politicians want a blackout? In the abstract, everyone is for higher turnout, but not many people care so much about turnout that they will change the law. There has to be some more pressing reason, and that reason has to do with strategic voting. If I remember correctly, the blackout period was introduced in the late 1990s by legislators who were sick of hearing that people shouldn’t vote for them because they were sure to win or sure to lose. We tend to think of strategic voting as something that mostly affects third party candidates, but remember that under the old SNTV electoral system with multiple seats in each district, KMT and DPP legislators also had to worry about polls that said they were sure winners or sure losers. Either way could spell disaster. Too much information could throw an election into utter chaos (and this eventually happened in the heavily polled 2004 legislative elections). In short, practical politicians wanted voters to vote sincerely (and let the strongest candidates win).
It might be argued that the state has a compelling interest in sincere voting as well since democratic elections are supposed to reflect popular opinion. However, I think that argument is not as clean as the turnout argument. In fact, I don’t think there is anything undemocratic at all about encouraging interest aggregation at a lower level (the voter) rather than at a higher level (the legislature).
A side note on the xfutures market. The CEC ordered the xfutures market to close during the blackout period, and there is some discussion about the legality of this decision. I think the CEC took an entirely reasonable position. While xfutures is not technically doing surveys, they have marketed their product as serving the same predictive function as surveys, except they believe they are more accurate. Since the purpose of the law is to increase uncertainty in the days before the election, extending the blackout to xfutures is certainly within the spirit of the law. Moreover, you could argue that xfutures really is a version of a public opinion survey. They are aggregating large numbers of opinions to produce an estimate of the entire population’s behavior. Xfutures is not relying on qualitative analysis or expert opinion; it is asking ordinary people. Their sampling methods and weighting methods are different, but you could argue that they are engaged in the same basic enterprise. Random sampling doesn’t seem to be the necessary criterion for inclusion in the blackout. If I’m not mistaken, the blackout extends to internet surveys, and they certainly don’t use random sampling. (I can’t find any of those “Who will you vote for?” polls on the big three newspaper sites.)
Who will the blackout benefit? I think the blackout benefits Soong and, indirectly and to a much lesser extent, Tsai. The purpose of the blackout is to increase uncertainty. Something might happen that Soong can claim has fundamentally changed the race since the last polls. It doesn’t even have to be that convincing; Soong simply needs to give people who really want to vote for him a rationalization so that they can do what they want to do. At the very least, Soong doesn’t have to answer questions every day about the most recent poll showing him at 6% or some other awful number. He can turn the focus elsewhere, which I’m sure is what he would prefer to do. Moreover, the KMT can’t legally bring polling numbers back into the conversation. They have to use vague statements that are not backed up by hard data.
To the extent that Soong’s voters are predominantly drawn from the blue camp pool of votes, this is good news for Tsai. I think the effect is probably limited. Many of Soong’s supporters would simply stay home if they didn’t vote for Soong. Perhaps half would vote for one of the other two candidates and two thirds of those would vote for Ma. Since the final public polls had Soong somewhere around 6%, that works out to about a 1% net gain for Ma. If something happens during the blackout to convince more of those voters to stay with Soong, the 1% shrinks. This is a close election and 1% might be the difference, but other factors will likely be more important.