Sabato's Crystal Ball

The Old College Try: Republican-Style

Larry J. Sabato, Director, U.Va. Center for Politics July 30th, 2003

facebooktwittergoogle_plusredditpinterestmail

In the June 2003 Crystal Ball, we manipulated the Electoral College to find a way for Democrats to compete, and potentially win, in November 2004. In this month’s Crystal Ball musing, we turn the tables. How should George W. Bush approach his old friend, the Electoral College? The president has not forgotten that the College – and the Supreme Court – handed him the Oval Office, despite a popular vote deficit of almost 540,000 votes. And to be sure, Bush, Karl Rove and company are carefully examining every potential combination of states to deliver a second term, with or without a popular plurality.

The White House does not share its calculations with us, alas, but the Crystal Ball will attempt to conjure up the best formula for a Bush re-election strategy. With political science as our training, it will surprise few readers that we’ll approach this task using a statistical twist.

The easy, and somewhat misleading, way to accomplish our goal would be simply to look at the 2000 election numbers to see where Bush was strong, weak, and on the margin. Naturally, that has to a part of the process, but it ignores the fact that all states evolve politically over time – some quickly, some slowly, some in a Democratic direction, some with a Republican tilt, a few others de-aligning from the two major parties (think Minnesota, for instance). So how can we tell what’s happening in each state? All measures are imperfect, but the best indicators we have are those created by the voters in the elections held in 2000 and 2002. (Yes, for simplicity’s sake, we are eliminating off-year and special elections. Attention purists: Please forgive us.) The 2000 presidential contest remains our guiding light, so let’s create a scale that reflects the state-by-state vote:

Now for the more recent 2002 elections… All the contests were sub-presidential, and involved their own peculiarities, incumbencies, candidate gaffes, and the like. Therefore, we can presume that while they tell us something about the “state of affairs” in each state, they cannot be as revealing as the 2000 vote. The point scale must be appropriately diminished then. So for each state, let’s give 3 points to each state that elected a Republican governor in 2002, and subtract 3 points from each state that voted in a Democratic governor. Similarly, let’s give or take 3 points for each Republican or Democratic U.S. senator elected in 2002. Now, let’s also add or subtract just 1 point for each Republican or Democratic House member elected in 2002.

That wasn’t so arduous, was it? Take Georgia, the GOP’s new favorite state. Bush carried Georgia by more than 10 percent in 2000 (11.65 percent to be exact), so the state starts out +10. In 2002, a Republican governor won, adding 3 points; a Republican senator won, adding another 3 points; and the U.S. House delegation elected was 8R, 5D, for a net plus of yet another 3 points. The grand total for Georgia, then, is +19, a near-lock for George W. Bush in 2004.

All the other states’ scores can be viewed here. Notice that the states are listed in the order of Bush’s popular percentage in 2000, from Wyoming’s gargantuan 67.76 percent (a victory of over 40 percentage points in Dick Cheney’s home state) to Rhode Island’s pitiful 31.91 percent (a drubbing of over 29 percentage points). Yes, we are ignoring the Democratic island of D.C., where the Gore landslide of 85.16 percent to Bush’s 8.95 percent is all you really need to know.

Before the analysis begins, one final important note: We have left Ralph Nader out of the analysis even though he unquestionably matters. Without him, Al Gore would be president today (look at the Florida and New Hampshire numbers if you doubt the Crystal Ball). Should Nader or another Green Party candidate be nominated for the White House in 2004, our state numbers will probably hold up fairly well. But should the Greens surprise us and endorse the Democratic candidate, or just not run anybody at all, then the Democrats will gain vital points in swing states, especially Florida, Iowa, Maine, Michigan, Minnesota, Nevada, New Hampshire, New Mexico, Ohio, Oregon, Pennsylvania, Washington, and Wisconsin. We’ll revisit our calculations if the unexpected happens and the Greens opt out.

According to the scores achieved by the 50 states in the Crystal Ball’s analysis, Bush’s absolutely solid states, unavailable to the Democrats except in the unlikely case of a 1964-style landslide, are: Alabama-9, Alaska-3, Georgia-15, Idaho-4, Indiana-11, Kansas-6, Kentucky-8, Mississippi-6, Montana-3, Nebraska-5, North Carolina-15, North Dakota-3, Oklahoma-7, South Carolina-8, South Dakota-3, Texas-34, Utah-5, Virginia-13, Wyoming-3. The total electoral votes available in these states are 161.

Likely Bush states, but ones where some time and money must still be spent to guarantee victory, include: Arizona-10, Arkansas-6, Colorado-16, Louisiana-9, Nevada-5, New Hampshire-4, Ohio-20, Missouri-11, Tennessee-11, and of course, Florida-27. Recent electoral trends in Missouri, New Hampshire, Ohio, and Florida have made these states more secure for Bush than in 2000 – though not impregnable. Florida, for instance, would obviously be far more competitive should Sen. Bob Graham be on the Democratic ticket in either position. The total electoral votes available in these states are 112.

As long as Bush carries every single state in these two categories, he will win, barely once again, with 273 electoral votes. There is no margin for error here, though. And the margin of victory comes entirely from the Census redistribution of House seats (and electoral votes) in 2001. Remember, Bush’s 271 total from 2000, one more than the required 270, is now 278 due to reapportionment.

Notice anything missing? That’s right, it’s West Virginia, with five electoral votes. One of the three biggest Republican surprises of election night 2000 (the others were Bush’s carrying of Tennessee and Arkansas), heavily Democratic West Virginia was the most unlikely of places for a GOP breakthrough. And according to our analysis, it is the single most threatened Bush state for 2004, with a Bush point score of just 3. The Democrats will pour money and lavish attention on the Mountain State in 2004, as will Bush. He’ll have to do so if he hopes to keep it in his column. Scandal-plagued Democratic Gov. Bob Wise will make it easier on Bush if he’s on the November ballot again, but that seems increasingly unlikely.

Now things get interesting. There were nine states, considered highly competitive before Election Day 2000, that George Bush lost by margins ranging from a few tenths of a percent to about 5 percentage points. In order of the percentage loss in 2000, they are: New Mexico (0.06 percent), Wisconsin (0.22 percent), Iowa (0.32 percent ), Oregon (0.44 percent), Minnesota (2.41 percent ), Pennsylvania (4.17 percent ), Maine (5.12 percent ), Michigan (5.14 percent), Washington (5.58 percent).

However, the Crystal Ball’s analysis suggests that the order of targets has changed, and that Bush’s best shot for a pick-up among these states may be Minnesota (+1), especially after the strong GOP victories in 2002. The Republican target states following Minnesota appear in this new order: Iowa (-6), New Mexico (-2), Oregon (-6), Pennsylvania (-3), Wisconsin (-6). Further out of reach are: Maine (-9), Michigan (-10), and Washington (-10). Bush will certainly want to capture at least a few of these states to avenge 2000, and their total of 92 electoral votes are an essential cache should he lose any state in the “likely” category.

Again to no one’s surprise, Bush can forget about many other states, unless conditions favor a landslide or near-landslide. (A landslide is defined as 55 percent or more of the popular vote, and usually results in a candidate’s carrying more than 40 states.) In this solid Democratic category we find: Connecticut-7, D.C.-3, Delaware-3, California-55, Hawaii-4, Illinois-21, Maryland-10, Massachusetts-12, New Jersey-15, New York-31, Rhode Island-4 and Vermont-3. The electoral vote total here is 168 votes.

Somewhat incredibly, the White House political operation keeps making noises about contesting New York (-16), New Jersey (-14), and the state with the highest anti-GOP score, even more than Massachusetts and Rhode Island – California (-25). Yes, perhaps the memories of Sept. 11 and the New York-based 2004 Republican National Convention will make New York and New Jersey competitive, and maybe if the extremely unpopular Gov. Gray Davis survives the recall and lives to be hated for the rest of his term, California can be won. The Crystal Ball wouldn’t want to wager the milk money on any of this, however – barring that speculative landslide.

Well, there you have it, dear Crystal Ball readers. In two monthly updates, close to a year and a half before the November election of 2004, we have laid out competing plans for the Electoral College. Both parties will give it the old college try, but only one will be admitted and graduate to the presidency for four years. Even your dependable Crystal Ball can only guess at this point which one it will be.