Sabatos Crystal Ball

History’s Paragraph for the 2006 Election

Lengthy and dramatic or brief and underwhelming?

Larry J. Sabato, Director, U.Va. Center for Politics June 1st, 2006


It’s humbling for all involved in America’s electoral process to realize that each midterm election season–all the contests put together–comprises no more than a paragraph in the history books. The significant elections merit a bold, detailed paragraph, while the run-of-the-mill midterms get a tepid, sketchy paragraph. Most election paragraphs are tepid; big midterm earthquakes are rare.

So what will 2006’s paragraph look like? Will it be lengthy and dramatic, or brief and underwhelming?

To attempt to answer this, let’s see if we can discern any useful historical patterns among the post-World War II midterm congressional elections. Here’s my attempt to write an appropriate paragraph for the fifteen midterms from 1946 to 2002, from Midterm Madness (Rowman & Littlefield, 2003):

It’s obvious from these descriptions that the bold midterms that really mattered–the ones that dramatically affected national policy and politics–were 1946, 1958, 1966, 1974, 1994 and 2002. In the cases of 1958, 19 66 and 1974, the midterms served as a prelude to a White House takeover by the opposition party. In 1946 the results were misleading, because the New Deal coalition of FDR was still strong–strong enough to enable Harry Truman to win an upset victory in 1 948. In 1994 the GOP landslide probably helped to reelect President Clinton two years later, ironically. And in 2002, President Bush’s midterm success signaled he would have a good chance to be reelected, as he was in 2004.

Only about a third of the midterms have made the cut into the “significant” category. For our purposes here, we’re also going to add three other midterms that had a moderate but telling impact on American politics: 1950, 1982, and 1986. This trio will further enrich our analysis and predictions.

What makes a midterm historically memorable? All of these elections involved two or more of the following critical factors:

Which factors played well in the nine modern midterm elections we’ve chosen to examine leading up to 2006?

Election Year Presidential Popularity War Economy Scandal Social Issues
1946 X (low) X
1950 X (low) X X X
1958 X X
1966 X (low) X
1974 X (low) X X
1982 X (low) X
1986 X
1994 X (low) X X X
2002 X (high) X
2006* X (low) X X X

* as of June

The most consistent indicator of a forthcoming “significant midterm” has been presidential popularity, almost always the lack of it. In 1946, 1950, 1966, 1974, 1982 and 1994, the incumbent President’s last Gallup Pol l rating prior to the election (usually late October to early November) was in the 30s to mid-40s, with the exception of new President Gerald Ford, who was still in his brief honeymoon despite his pardon of President Nixon. (Nixon’s August resignation day rating of 23 percent was more relevant to the election of 1974 than Ford’s rating. The midterm was something of a public judgment on the Watergate scandal.) Here are the pre-election ratings for all nine midterms we are examining:

President Poll Dates Poll Approve Disapprove No Opinion
Truman Sep. 13-18, 1946 Gallup 33 52 15
Truman Oct. 20-25, 1950 Gallup 41 46 14
Eisenhower Oct. 15-20, 1958 Gallup 57 27 16
Johnson Oct. 21-26, 1966 Gallup 44 41 15
Ford Oct. 18-21, 1974 Gallup 54 29 18
Reagan Oct. 15-18, 1982 Gallup 42 48 10
Reagan Oct. 24-27. 1986 Gallup 63 29 8
Clinton Nov. 2-6, 1994 Gallup 46 46 8
Bush Oct. 31-Nov. 3, 2002 Gallup 63 29 8

It is vital to note the exceptions to the “presidential popularity” midterm indicator. In 1958 Eisenhower was able to maintain his personal approval rating at a remarkable 57 percent, but it couldn’t help his party in his sixth-year itch election. Similarly, in 1986 Reagan had a robust 63 percent (just prior to the Iran-Contra scandal, which broke several weeks after the election), yet the Democrats grabbed the U.S. Senate anyway. The most unusual exception to the p opularity guideline occurred in 2002, when George W. Bush had a Reagan-like 63 percent approval rating. Unlike Reagan, though, he was able to transfer some of his post-September 11th popularity to the GOP in capturing the Senate and adding House seats. Of course, this was his first midterm election, not the sixth-year contests.

War has mattered three times before 2006. The controversial Korean War in 1950 and the deeply unpopular Vietnam War in 1966 cost two Democratic Presidents, Truman and LBJ, many congressional seats. Conversely, the hi ghly successful Afghanistan War in 2001 and the overall War on Terrorism, which had strong public backing, helped Bush and the Republicans in 2002.

The pocketbook issue, generated by bad economies, damaged the President’s party on six occasions–1946 (when the transition from a wartime economy to a peacetime economy proved very difficult), 1950, 1958, 1974, 1982 and 1994. The economy wasn’t good in 2002 either, but voters chose to overlook it or attribute it to the effects of September 11th.

Scandal has played a role four times, in 1950, 1958, 1974 and 1994. With the exception of Watergate, a mega-scandal, in 1974, corruption has been more a supplementary than a central explanation for the election resul ts.

Key social or domestic issues have been factors twice, in the Democrats’ Senate triumph of 1986, where they used fears of possible GOP cuts in Social Security to win over senior citizens, and in 1994 when the Republi cans turned the tables on the Democrats by emphasizing cultural concerns such as abortion, gun control, and gay rights.

Now that we have an enlightening historical perspective as a template, let’s see how we can project the 2006 midterm contests. The omens are dark for Republicans, at least as they appear in June. President Bush’s app roval rating hovers in the low-to-mid 30s; Bush needs a sizeable recovery quickly, because a modest boost in the polls won’t help all that much at his cellar level. The Iraq War could hardly be less liked by the voters, or more directly tied to the GOP–a lready on a par with Korea, and increasingly resembling Vietnam, though without the draft or the jungle. In addition, there are scandals galore, though as we have argued in a previous recent analysis, Democratic congressional scofflaws prevent this factor from being an unalloyed advantage for Democrats. (Since that piece was written, Congress man Bill Jefferson of Louisiana has attracted a great deal of unwanted publicity that can only damage Democratic claims of relative purity.) Perhaps the coming Abramoff lobbying indictments will re-energize scandal and tilt it heavily against the Republic ans, assuming the targets are overwhelmingly GOP.

As far as social issues go, there are already anti-gay marriage amendments on the November ballot in six states (Idaho, South Carolina, South Dakota, Tennessee, Virginia, and Wisconsin), with probably more to join in . This subject tends to motivate a disproportionately conservative electorate, and so it likely will be a plus for Republicans, just as it was in the eleven states that featured similar referenda in 2004. But other social issue s that have emerged in 2006, immigration and stem-cell research, may assist Democrats in some places. For instance, the fierce debate over immigration reform could possibly shift more of the critical Hispanic and Latino swing vote from the GOP to the Democrats.

Only a generally robust, stable economy can fairly be counted as a Republican plus in 2006. However, most voters do not yet perceive the economy to be nearly as solid as it actually is–a by-product of gloom over Ira q, dislike of President Bush, unhappiness over skyrocketing gasoline prices, and troubles in industries such as autos and the airlines. If Republicans hope to minimize their losses, they must focus more on this crucial election category. Without credit fo r the economy, voters may well conclude that “everything is going to hell in a hand-basket,” a conclusion that almost always leads the electorate to sing a chorus of “It’s time for a change”.

With five months to go before November, can every piece of this analysis be transformed? It’s certainly possible, but it’s also not very likely. The problems bedeviling Bush and the GOP may get a bit better or a little worse, b ut they appear to be intrinsic to this election year. Therefore, the only real question is how many seats the Democrats will gain in the Congress and the statehouses, not if they will gain.

Take the case of the U.S. House of Representatives. It is difficult to see the Democrats picking up fewer than five seats net–a third of the fifteen seats the party needs for a House takeover. Most election rating sites, including the Crystal Ball, have been gradually upgrading their estimates of likely Democratic gains, with the number now ranging from seven to twelve seats. That is approaching the magic level of addition required for a House majority. At the same time, early projections are not reality and plenty of events could alter the landscape just enough to preserve Republican control of the House.

Look again at our table of post-World War II historical precedents. There is cold comfort there for the GOP, and at least preliminary cheer for Democrats. The election paragraph for 2006 is not yet written, but it will be a sur prise if its content does not please Democrats more than Republicans. But will the results make much difference for governance by giving real power to the Democrats in at least one house of Congress? That is the key question that can only be an swered by the events to come, the campaigns yet to be run, and the final decisions of the voters–the ones who, far more than scholars, actually write the midterm election passage in the book of American history.