CHAMPAIGN, Ill. - While national preference polls are shedding little light on the outcome of the 2004 presidential race, state-level polls and electoral history can yield important insights, including "some unsettling possible outcomes in next month's election."
So says Peter F. Nardulli, the author of a new study that is based on 20 years of research on state-level presidential voting patterns in the United States between 1828 and 2000. According to Nardulli, "One of the most important insights derived from joining electoral history and current state polls is that there is a non-trivial possibility of a tie in the Electoral College."
The possibility of a tie in the Electoral College is unsettling because it could lead to "a legitimacy crisis that would dwarf the one that emerged in Florida in the election of 2004," Nardulli said. Under the Constitution, the House of Representatives, voting as state delegations with one vote allocated to each state, would choose the president in the event of such a tie.
"This archaic, pre-democratic mechanism, when joined with the complex nature of the American federal system, would almost certainly lead to a situation in which a sizeable number of state delegations would overturn the winner of the state's popular vote," Nardulli said. "The House vote could also conceivably award the presidency to a candidate that did not win the national vote."
This is "not a pretty scenario for the world's leading democracy at the dawn of the 21st century," said Nardulli, who has updated his published forecast for the 2004 election. Still, he said, "The legitimacy crisis a tie vote in the Electoral College would cause is wholly avoidable."
For one thing, there is "nothing but party discipline and career ambition" that mandate members of Congress to vote along party lines, he said.
"Moreover, this potential legitimacy crisis can be avoided without the Republicans giving up what almost would be a certain House victory."
Nardulli is calling for a resolution that is "consistent with both the U.S. Constitution and contemporary democratic norms": that candidates for seats in the House - all of whom are currently standing for election - discuss the possibility of a tie in the Electoral College with their constituents and announce, before the election, how they would vote in the case of a tie. Newspaper and civic leaders should insist upon such a public statement, particularly in the states most likely to be affected.
"Candidates for House seats could make defensible arguments to vote for the candidate that their party nominated; that wins the district's popular vote; wins the state's popular vote; or wins the nation's popular vote.
"What is essential is that these arguments are presented - and firm commitments made - during the campaign," Nardulli said.
Nardulli's insights come as a result of having updated an analysis he first published in the October issue of PS (Political Science & Politics), a publication of the American Political Science Association. Nardulli is a voting expert who pioneered a "normal-vote approach" to capturing changes in presidential voting patterns.
According to Nardulli a professor of political science at the University of Illinois at Urbana-Champaign, a state's normal vote in a particular election refers to what the margin of victory in the election would be if citizens' habitual voting patterns - voting Democratic, voting Republican, abstaining from voting - solely determined the vote.
"In a sense, the normal vote measures the relative size of the major parties' electoral base in a state," Nardulli said, noting that the normal-vote procedure he uses derives an estimate of the normal vote by calculating "five-election moving averages."
Nardulli's initial forecasts showed that the Democrats began the 2004 campaign with "meaningful normal-vote advantages" in states with a total of 282 Electoral College votes; the Republicans had meaningful advantages in states with 182 votes.
He updated these analyses using available state polls for 17 states that were initially viewed as "battleground" states. These polls show that "Republicans are in no danger of losing any state in which they have a normal-vote advantage."
"Moreover, available state polls suggest that the Republicans are in a good position to pick up one competitive state, Louisiana, and two states in the Democratic fold, Arizona and West Virginia," giving them a revised Electoral College total of 202.
In addition to losing Arizona and West Virginia to the Republican fold, available state polls reveal that four states in which the Democrats have normal-vote advantages are "extremely competitive," he said: Iowa, Missouri, New Mexico and Wisconsin. Polls also reveal that the Democrats are in a good position to pick up New Hampshire, which gives them a revised Electoral College total of 242.
Using this revised forecast, he calculated an "exhaustive listing of outcome scenarios by systematically varying the outcomes for the competitive states." This exhaustive list of outcome scenarios was then joined with data on Electoral College votes to identify winners, losers and tie scenarios. He found:
• If win scenarios for the 11 currently most-competitive states are examined, 11 tie scenarios exist; 29 tie scenarios exist if the 13 most-competitive states are examined. These tie scenarios are based on the assumption that the Colorado ballot initiative to allocate Electoral College votes proportionally either fails or is struck down by the courts.
"Perhaps most likely of these scenarios is the one in which the Republicans win all of the competitive states leaning toward them and pick up Arizona and Missouri, while the Democrats pick up New Hampshire.
"The Democrats are currently ahead in the Granite State, the Republicans are ahead in Arizona, and the Show-Me-State is currently competitive but learning toward the Republicans."
• In the case of a tie in the Electoral College, assuming that state delegations voted along party lines - "a very safe bet," Nardulli said - the Republicans would win the presidency.
They have majorities in 30 states; the Democrats have majorities in 16 states. Four states have evenly balanced delegations, potentially making them unable to cast a ballot in the House election.
"This is unsettling from the perspective of democratic theory and practice," Nardulli said, "because the Democrats could win a plurality of the votes cast in the nation and again lose the election, even though it would be wholly constitutional."
• Compounding the difficulties a tie scenario would cause is the fact that partisan House voting would inevitably lead to a larger number of instances in which a state's delegation would overturn the popular vote.
"For example, if the tie scenario mentioned previously - New Hampshire goes Democratic, Arizona and Missouri go Republican, and all other states vote in accord with their normal vote - there will be eight states in which the Democrats would win the popular votes but lose in a partisan House vote." Those eight states are Connecticut, Delaware, Iowa, Illinois, Michigan, New Hampshire, New Mexico and Pennsylvania.
"Because of evenly balanced state delegations," Nardulli said, "partisan voting in the House election would negate Democratic popular victories in three other
states - Minnesota, Vermont and Wisconsin."
There will be five states in which the Republicans would win the popular votes but lose in a partisan House vote - Arizona, North Dakota, South Dakota, Tennessee and Texas.
"Because of evenly balanced state delegations, partisan voting in the House election would negate a Republican victory in Mississippi."
Details on the process Nardulli used to derive normal-vote estimates are available online in Appendix II.Graphs of all state normal-voting patterns are available for review in "State Level Voting Patterns for President, 1828-2000," which is accessible at the same Web address.
Tables for his updated forecast and listing of tie scenarios can be found at the same address under "Revised 2004 Electoral College Forecast."