Should Election Day Be a Holiday?

A simple, practical step might summon the ranks of nonvoters from the civic void

REGARDLESS of who wins in next month's midterm elections, a sure bet is that less than half of the voting-age population will actually participate. The percentage of the electorate casting ballots for the House of Representatives has fluctuated between 33 and 45 percent over the past sixteen midterm elections. Recent turnout rates suggest that the percentage in 1998 will probably be near the bottom of this range, and quite possibly even lower. In 1996 the presidential-election turnout fell below 50 percent for the first time since the early 1920s -- when women had just received the franchise and had not yet begun to use it as frequently as men. Last year not a single one of the eleven states that called their citizens to the polls managed to get a majority to vote. The best turnout occurred in Oregon, where a heated campaign debate had taken place on the question of whether to repeal the state's "right to die" law. The worst turnout last year was a shockingly low five percent, for a special election in Texas. This occurred even though Governor George W. Bush stumped the state for a week, urging people to participate and promising that a "yes" vote would result in a major tax cut.

Universal suffrage means that everyone should have an equal opportunity to vote, regardless of social background. But over the past three decades studies have found increasing biases in turnout. In particular, people without college degrees have become less likely to go to the polls. Statistics from the Census Bureau on turnout by educational achievement make the point. Respondents were asked if they had taken part in the most recent national election.

Since 1966 turnout rates have declined most sharply among people at the lower levels of education. In 1994 people with no college education made up 53 percent of the adult population but only 42 percent of the voters.

Turnout is now also greatly related to experience in life. Turnout rates have always been lowest among young people; perhaps this is why there was relatively little opposition in the early 1970s to lowering the voting age to eighteen. But not even the most pessimistic analysts could have foreseen the record-low participation rates of Generation X, as shown in the following census findings on age and turnout:

The low turnout among young voters today is paradoxical given that they are one of the best-educated generations in American history. Even those who have made it to college are expressing remarkably little concern for politics. Chelsea Clinton's class of 2001 recently set a new record for political apathy among college freshmen: only 27 percent said that keeping up with politics was an important priority for them, as opposed to 58 percent of the class of 1970, with whom Bill and Hillary Clinton attended college.

Of course, Chelsea's classmates have not seen government encroach on their lives as it did on the lives of their parents -- through the Vietnam War and the draft. Nor has any policy affected them as directly as Medicare has affected their grandparents. It is noteworthy that senior citizens are actually voting at higher rates today than when Medicare was first starting up. Political scientists used to write that the frailties of old age led to a decline in turnout after age sixty; now such a decline occurs only after eighty. The greater access of today's seniors to medical care must surely be given some credit for this change. Who says that politics doesn't make a difference?

Yet it is difficult to persuade people who have channel surfed all their lives that politics really does matter. Chelsea's generation is the first in the age of television to grow up with narrowcasting rather than broadcasting. When CBS, NBC, and ABC dominated the airwaves, their blanket coverage of presidential speeches, political conventions, and presidential debates sometimes left little else to watch on TV. But as channels have proliferated, it has become much easier to avoid exposure to politics altogether. Whereas President Richard Nixon got an average rating of 50 for his televised addresses to the nation, President Clinton averaged only about 30 in his first term. Political conventions, which once received more TV coverage than the Summer Olympics, have been relegated to an hour per night and draw abysmal ratings. In sum, young people today have never known a time when most citizens paid attention to major political events. As a result, most of them have yet to get into the habit of voting.

The revolutionary expansion of channels and Web sites anticipated in the near future is likely to worsen this state of affairs, especially for today's youth. Political junkies will certainly find more political information available than ever before, but with so many outlets for so many specific interests, it will also be extraordinarily easy to avoid public-affairs news altogether. The result could well be further inequality of political information, with avid followers of politics becoming ever more knowledgeable while the rest of the public slips deeper into political apathy. This year's expected low turnout may not be the bottom of the barrel.

SOME commentators welcome, rather than fear, the decline in turnout rates in America. If people do not vote, they say, citizens must be satisfied with the government. There is a certain logic to this view, because if nonvoters were extremely disgruntled with our leaders, they would undoubtedly take some political action. However, to argue that nonvoters are content with government just because they aren't actively opposing it stretches the logic too far. When the 1996 National Election Study asked people to rate their satisfaction with how democracy works in the United States, nonvoters were less positive than voters. Furthermore, young people were more than twice as likely as senior citizens to be dissatisfied with American democracy.

Why should young adults be satisfied with government, given how few benefits they receive from it in comparison with their grandparents? But until they start showing up in greater numbers at the polls, there will be little incentive for politicians to focus on programs that will help the young. Why should politicians worry about nonvoters any more than the makers of denture cream worry about people with healthy teeth? It is probably more than coincidental that Clinton's two most visible policy failures -- the 1993 economic-stimulus package and the 1994 effort to establish universal health care -- had their strongest backing from people who were not even registered to vote. Congressional Republicans may rationally have anticipated that many of these proposals' supporters were unlikely to be judging them in the 1994 elections.

After the Republican takeover of Congress in 1994, I saw a bumper sticker that read NEWT HAPPENS WHEN ONLY 37 PERCENT OF AMERICANS VOTE. Although I don't usually let bumper stickers determine my research agenda, this one piqued my interest. Would the Republicans have won the majority of House seats if turnout had been greater? A simple way to address this question is to assess how much difference it would have made if voters had mirrored the adult population in terms of education. According to the 1994 National Election Study, 30 percent of voters who lacked a high school diploma and 62 percent of voters with college degrees voted for Republican candidates for the House. Increasing turnout among the least educated citizens would thus have made some difference. If turnout rates had been equal in all education categories, the Republican share of the vote would have fallen from 52.0 to 49.2 percent.

Although it is unlikely that people of differing education levels would ever vote at exactly the same rate, this is only one of many biases in electoral participation. A more comprehensive method of estimating the impact of higher turnout is to gauge the attitudes of nonvoters toward those factors that influenced voters in 1994: party identification, approval of Clinton, stands on issues, and incumbency. Examining only survey respondents who were registered but did not vote, I found that these nonvoters would have favored Democratic candidates by an even greater margin than that by which actual voters supported the Republicans. Had all registered citizens gone to the polls, the Republicans' share of the vote would have been reduced by 2.8 percent -- exactly the same estimate as arrived at above. If this loss occurred in all districts, the Republicans would have won only 206 seats -- twenty-four fewer than they actually won, and twelve short of a majority.

Such findings, regrettably, suggest that nothing will be done to increase turnout in America. Few Republicans will want to correct a situation that has benefited them in the past. Yet until something is done, the House of Representatives will be representative not of the electorate but only of the minority that actually votes.

WHAT can be done to reverse the decline in turnout? At his first press conference after the 1996 election Bill Clinton was asked about the poor turnout and how to increase participation in the future. The President stumbled over this question -- he didn't really have an opinion on what could be done, and he concluded by asking the members of the press corps whether they had any ideas. Clinton's apparent frustration in addressing the question probably stems from his involvement in passing the 1993 Motor Voter Act. He and many others believed that its voter-registration reforms would increase turnout. But although the registration rolls swelled in state after state prior to the 1996 election, the turnout rate fell dramatically on Election Day. (The Census Bureau, paradoxically, found fewer people in 1996 than in 1992 who said they were registered. Apparently, the Motor Voter procedures made registering so easy that many forgot they had placed their names on the voting ledgers.)

Had Clinton been better advised on this subject, he would not have expected turnout to increase simply because registering to vote had become easier. North Dakota has since 1951 not required people to register in order to vote, yet it has seen turnout in presidential elections decline by 22 percent since 1960. Minnesota and Wisconsin have allowed citizens to register on Election Day since the mid-1970s, but they have lower turnout rates today than when they had tighter registration laws. In short, not even the most lenient voter-registration procedures are the answer to the problem of low turnout.

Clinton is said occasionally to remark that solutions to most public-policy problems have already been found somewhere -- we just have to scan the horizons for them. This certainly applies to increasing turnout. Three possible changes stand out as particularly likely to get Americans to the polls -- though, unfortunately, their probable effectiveness is inversely related to the plausibility of their ever being enacted in the United States.

If in an ideal democracy everyone votes, people could simply be required to participate. This is how Australians reasoned when they instituted compulsory voting after their turnout rate fell to 58 percent in 1922. Since then the turnout in Australia has never fallen below 90 percent, even though the maximum fine for not voting is only about $30, and judges readily accept any reasonable excuse. However, American political culture is based on John Locke's views on individual rights, whereas Australian culture was shaped by Jeremy Bentham's concept of the greatest good for the greatest number. Most Americans would probably assert that they have an inviolable right not to vote.

Beyond that, it is debatable whether we really want to force turnout rates in America up to 90 percent. People with limited political knowledge might deal with being compelled to vote by making dozens of decisions in the same way they choose lottery numbers. In Australia this is known as the "donkey vote," for people who approach voting as if they were playing the old children's game. Given Australia's relatively simple electoral process, the donkey vote is a small proportion; in America it would probably be greater.

Evidence from around the world indicates that our turnout rates could be increased if we adopted some form of proportional representation. In our winner- take-all system many Americans rightly perceive that their votes are unlikely to affect election outcomes. Proportional representation changes this perception by awarding seats to small voting blocs. The threshold for representation varies by country, but typically any party that receives more than five percent of the national vote earns seats in the legislature. Almost inevitably when proportional representation is instituted, the number of political parties grows. And with a range of viable parties to choose from, people tend to feel that their choice truly embodies their specific interests. Hence they are more likely to vote.

If we were to adopt proportional representation, new parties would be likely to spring up to represent the interests of groups such as African-Americans, Latinos, and the new Christian right. Although this would give members of these groups more incentive to vote, and thus would raise the low turnout rates of minority groups, a price would be paid for this benefit. The current system brings diverse groups together under the umbrellas of two heterogeneous parties; a multi-party system would set America's social groups apart from one another. Proportional representation therefore seems no more practical on the American scene than compulsory voting.

A simple but effective change, however, could be made in election timing. An ordinary act of Congress could move Election Day to a Saturday or make it a holiday, thereby giving more people more time to vote. An 1872 law established the first Tuesday after the first Monday in November as Election Day. At that point in history it made little difference whether elections were on Saturday or Tuesday, because most people worked on Saturday. Only Sunday would have been a day free of work, but with elections in the late nineteenth century being occasions for drinking and gambling, that option was out of the question in such a religious country.

Americans have become quite accustomed to Tuesday elections, just as they have to the nonmetric system for weights and measures and other artifacts of another time. State after state has set primary-election dates on Tuesdays -- all twentieth-century decisions, some of them quite recent. It would be difficult to change this custom. Furthermore, there would probably be some resistance from religious minorities that observe the sabbath on Saturday.

An alternative would be to declare Election Day a national holiday. This would probably be resisted on the basis of cost. A solution would be to move Election Day to the second Tuesday of November and combine it with Veterans' Day, traditionally celebrated on November 11. This would send a strong signal about the importance our country attaches to voting. And what better way could there be to honor those who fought for democratic rights than for Americans to vote on what could become known as Veterans' Democracy Day?