I have always wondered why so many people are concerned about voter turnout and and so few are concerned about the fact that most of the voters know little or nothing about what they're voting about.
The state of Arizona has established a $1 million dollar lottery for the purpose of trying to increase voter turnout. Each voter will, in effect get a lottery ticket that gives him or her a chance of winning. I don't know whether the Arizona plan will increase turnout or not. In the last off-year election in 2002, some 1.225 million Arizonans voted. If the same number turns out in 2006, this means that the expected winnings of each voter will be about $1 (1/1.225 million or possibly 2/1.225 million if there is a separate jackpot for primaries). Doesn't sound like much of an incentive to me, though perhaps some people might come just because they enjoy participating in lotteries.
The more important question, to my mind, is why we should care about turnout levels at all. Obviously, a nation where only 1% of the people vote is likely to have serious problems because the voting population will be highly unrepresentative of the public as a whole. However, in the real world, we are not talking about the difference between 1% and 80% turnout, but about the difference between, say, 50% and 70% at most. In theory, even an electorate with 50-60% turnout (roughly the US average in recent presidential elections) might be highly unrepresentative. However, numerous studies show that, at least in the US, the partisan and ideological affiliations of nonvoters are roughly similar to those of voters (see, e.g., Democratic Party election consultant Ruy Teixeira's important book, The Disappearing American Voter (1992)). The 2004 election, which saw an increase in turnout from about 51% in 2000 to 59%, is not an exception to this rule. To make a long story that I cannot fully document here short, the Republican turnout operation essentially offset the Democratic one and vice versa.
I follow the academic literature in this field fairly carefully and I have yet to come across a study that provides any evidence that countries such as Australia and Belgium, which have turnout rates over 80% as a result of compulsory voting laws, have better public policy outcomes or stronger democratic control of government as a result. On the other hand, one of the few advanced nations with turnout levels lower than the US is Switzerland, which has turnout of about 33-40 percent in its national elections (comparable to US off-year elections and much lower than US turnout in presidential election years). Switzerland, of course, is usually considered one of the best-governed nations in the world. I am NOT claiming that this success was caused by low voter turnout, but low turnout certainly has not inhibited it.
Efforts to raise turnout cost money, including but not limited to Arizona's $1 million. If increasing turnout within the parameters that are relevant in the real world does not produce any significant benefit to society and does not change electoral outcomes, those resources might be better expended elsewhere.
Moreover, the non-problem of low turnout diverts scarce media and academic attention away from what I have repeatedly argued is the far more significant problem of widespread political ignorance (see, e.g., here). I don't worry about whether turnout in this year's election will be 40% or 50%; I do worry that most of the voters are likely to be ignorant of very basic facts about many of the candidates and issues.
UPDATE: My most recent article on the harm caused by widespread political ignorance can be found here.
UPDATE #2: Some commenters claim that nonvoters are disproportionately likely to be Democrats. This is a common claim in the public discourse, but it is contradicted by every serious study conducted by political scientists over the last 30 years or more. The Teixeira book I mention in the original post is just one of many such analyses. I should also note that the claim seems to have been falsified by the 2004 election, where there was a major increase in turnout (from 51 to 59%), but the Republican vote increased just as much as the Democratic vote, or even slightly more. If the (nonacademic) conventional wisdom was correct, we should expect a large increase in turnout (some 15% in this case) to benefit the Democrats, yet it strikingly failed to do so. Regular readers of this blog know that I don't have any great love for today's Republican Party, but it is simply not true that they stay in power because a disproportionate number of Democratic-leaning voters fail to turn out.