Before we plunge head-first into the election season, we should recount the various ways that midterm elections have affected the course of American politics over the years.

Of course, the most basic historical fact about midterm elections is that the president’s party almost always loses seats in the House, and loses seats in the Senate about two-thirds of the time. Even when the midterms have little broader impact—for example, on the next presidential election—they usually make life harder for the president. In typical years, he finds the composition of Congress less to his liking, both because fewer members of his party are there and because, often, the members of his party who were defeated were disproportionately drawn from those closely associated with the president or his ideas.

Less tangibly, when the opposition party gains seats, they gain—and are almost always declared by the media to have gained—an important (though mostly psychological) measure of political momentum. In a related vein, the modern plebiscitary presidency draws no small portion of its power from the perception of "mandate" flowing from the president’s election. Losses in the midterm year have the effect of negating that mandate, and perhaps even allowing the opposition to claim a counter-mandate.

Not to be ignored is the tendency of midterm elections to throw up a new set of attractive opposition leaders who, together with their party as a whole, test issues and themes for the next election.

When midterm elections follow the historical anti-administration pattern, they also have a long-term effect on American politics working from the bottom up. As it turns out, the midterm pattern manifests itself in state elections as well; the president’s party usually loses governorships (a real danger for Republicans this year) and state legislative seats. The effect on governorships can be particularly strong, since three-fourths of governors are elected in midterm years. Governors often run for Senate, and state legislatures are breeding grounds for future congressmen—meaning that the opposition party can plant seeds in midterm years that will bear great fruit, though sometimes many years later.

In some midterm years, the effect is even more pronounced. In about one-third of midterm elections since 1894, the result contributed significantly to a change in party control of the presidency two years later. Not only did the president’s opponents gain seats in Congress, but they gained control of the congressional agenda and used that control over the next two years to stymie and embarrass the president and his party. Sometimes (though not always) they also pursued an alternative policy approach, which attracted public attention and defined the party in a new and positive way for voters.

On the other hand, a few midterms, though inflicting serious losses on the president, had the effect of consolidating his position rather than upending it completely. In those cases, the midterms were an opportunity for voters to slow and refine a popular reform program without reversing it. The opposition cooperated by running a campaign aimed at making the president’s program more "efficient" or more "fair" rather than repealing it. The elections of 1938, 1982, and 1986 fit this mold. The president’s leverage was trimmed, but he or his party won the next presidential election and his reform program became a central fact of American political life.

And, of course, there are a handful of exceptions to the rule—cases in which the president’s party actually gained seats in the House or otherwise fought the opposition to a standstill. Prior to 1998, there had only been one House gain by a president’s party in the previous hundred years. Now there have been three; both 1998 and 2002 bucked the historical trend. One can also add 1962 to the list of exceptions, as John F. Kennedy’s Democrats gained a few seats in the Senate and lost only a very few in the House. In these cases, the benefits that typically accrue to the opposition party went to the president instead, ranging from a friendlier Congress to intangible momentum to a renewed sense of "mandate."

Altogether, midterm elections have often served as the starting and ending points of key policy eras. Woodrow Wilson’s progressive "New Freedom" era really began with the midterm election of 1910 and ended when Republicans regained control of Congress in 1918. The New Deal was presaged by the elections of 1930 and came to a screeching halt (as a program for additional reform) in 1938. The Great Society was foretold in the 1958 Democratic sweep, and hit a brick wall when Republicans made big gains in 1966. And the Reagan Revolution was anticipated by the tax-cutting fervor of 1978 Republicans, and could not be significantly added to after 1986, when Democrats regained control of the Senate.

Will 2006 be a normal midterm that takes seats from the president’s party and makes his life more difficult? Will it go further, serving as a platform for a Democratic recapture of the presidency in 2008? Will it be another exception, and if so, will the exception now be the rule after the third time in a row? What role will the midterm election play in the political and policy history of our time?

To bring light to these questions, I will be using my next several columns to examining in greater depth some of the key midterm elections of the last century.

Andrew E. Busch is a Professor of Government at Claremont McKenna College and an Adjunct Fellow of the Ashbrook Center.