Related Topics
election 2006

President Bush Job Approval

RCP Average
Send to a Friend | Print Article

2006: What About the Senate?

By Jay Cost

The recent election talk from our pundit class has been about the chances of a Democratic takeover of the House. Many expect this. I am open to this possibility; however, as I have written in the past, I think the data favors a Republican retention. I am also open to, and would be interested to read, a thorough argument that supports a Democratic takeover. Many pundits draw this conclusion, but I have yet to read the support necessary for it. In other words, my real objection is not so much to the conclusions of pundits, but to how they have drawn the conclusions. On the path to a declaration for the Democrats, they all stumble.

The biggest type of stumbling block is the systematic underestimation of evidence that cuts against their arguments. At best, many facts of importance, like 2004's 98.8% incumbent retention rate or 2006's incredibly low 4.6% incumbent retirement rate, are mentioned only to be unceremoniously dismissed. This is the sign of poor argumentation. It is not enough to proffer one's case by rallying the supporting facts. One must also handle the opposing facts.

One of these ignored items has to do with the Senate. It is, according to most, out of the Democrats' grasp. I strongly agree with this estimation. For the Democrats to take the Senate, they would have to defeat incumbents in Pennsylvania, Ohio, Montana, Missouri, and Rhode Island; win the open seat in Tennessee; and hold seats against strong challengers in Minnesota, Maryland, New Jersey and Washington. This amounts to a sweep of all 10 of National Journal's 10 most vulnerable races. Most would thus admit that the Senate is not on the table; those who make no such admission usually grow silent when asked to explain why they refuse.

The consensus on the Senate is actually a major problem for the consensus on the House. Historically speaking, the House only switches when the Senate switches. In other words, the improbability of a Democratic capture of the Senate is a sign that a capture of the House is improbable. Consider the following.

The 17th Amendment, which mandates the direct election of senators, took effect prior to the 1914 election. Since then, the Senate has changed hands 10 times due to the biannual congressional election. The House of Representatives, on the other hand, has only changed hands 6 times due to the biannual congressional election. (N.B. A 7th switch occurred in the middle of the 72nd Congress. The 1930 elections left the GOP with a slim majority. However, 14 representatives-elect died before the 72nd Congress convened, and the Democrats won enough of the subsequent special elections to take the House. This capture was "ratified" in the 1932 elections, which would have delivered Congress to the Democrats even if this tragedy had not occurred. So, let us henceforth identify 1932 as the 7th time that the House has switched since 1918.)

Furthermore, of these 7 times the House has switched, the Senate has also switched. Not only does the Senate switch more frequently, it always switches with the House. A switch in the Senate, therefore, seems to be a necessary, but insufficient, condition for a switch in the House. Conversely, a switch in the House is a sufficient, but not necessary, condition for a switch in the Senate. In other words, when the House has switched, the Senate has always followed; however, when the Senate has switched, the House has not necessarily followed. Thus, historically speaking, two scenarios are possible: the Senate alone changes or both the House and the Senate change.

Is this simply historical coincidence, or is a causal logic driving the correlation? A pattern that holds over 46 observations without exception is probably not random. Most important, however, is that this sort of pattern coheres with what we already know about Senate and House elections: namely, House elections are much less susceptible to national trends than Senate elections. This is the case for several reasons.

First, senators lack the ability to draw district lines to minimize opposing partisans. Second, Senate challengers tend to be more qualified and better funded than House challenges. Third, these challengers can use campaign resources more efficiently. Many House challengers cannot efficiently spend money on television advertisements because it is wasted on voters in other districts; however, with Senate elections, there are efficient ad markets. Fourth, senators are less able to cultivate close relationships with constituents - they lack the requisite geographical proximity. Fifth, senators are much more visible to the public; whereas House members can operate in the Capitol without much scrutiny, constituents tend to be more aware of their senators' activities.

Thus, Senate elections are contests where the partisan division is more equal, the average voter has a more balanced view of the candidates, and has more information about the issues in the race. Like House races, they tend to be referenda on incumbents; Senate incumbents are simply less favored. It is thus no surprise that senators' reelection rate is consistently lower than representatives'. It also no surprise that control of the Senate is more susceptible to change. If individual senators face enhanced competition, so does their partisan caucus.

Why, then, does the Senate always switch with the House? National political moods do not usually translate into changes in party control in the House. The reason for this is that individual members of the House are fairly invulnerable to that mood. However, they are not perfectly invulnerable. If the mood is sufficiently strong or sufficiently directed against one party, the House incumbency advantage is not enough. Since the House incumbency advantage is greater than the Senate advantage - we should expect the Senate to switch when the House switches. If the mood is strong enough to change the House, it will be strong enough to change the Senate. On the flip side, we can expect relatively milder political moods to change the more competitive Senate, but not the House.

This implies yet another point, always unmentioned among pundits inclined toward a Democratic takeover of the House. Contrary to the original intention of the Framers, the House is more insulated from the public than the President or the Senate. This is not because the House has altered its relationship with the public. It is because other branches have evolved while the House has remained the same. By original design, the Senate and the President were intended to be largely non-democratic institutions. However, by 1828 the President had become a fully democratic officer. The Senate became directly democratic in 1914. Elections for both evolved into kinds of referenda on the state of the union (in presidential elections, the contests are indeed referenda; in Senate elections, they are more like effective ones, thanks to the reduced incumbency advantage). This never happened with House elections. The House's relationship with the public has thus not changed much since the founding. It has never been a national assembly like the House of Commons. It has always been the meeting place of the representatives of different parts of the nation. It is democratic, but it remains democratic in the way that the Framers envisioned democracy in America.

Of course, one might respond, this argument could just as easily predict that both the House and the Senate will flip this year. The error that pundits are making, according to this line, is not with the House but with the Senate. Both are vulnerable. I do not find this compelling. One of the reasons pundits are so prone to write off the Senate is that they know more about the individual elections (this, by the way, is in keeping with senators' reduced incumbency advantage - individual senators are better known). They have a better sense of the electoral landscape, and therefore can appreciate that a net of six is prohibitively difficult. However, pundits know less of the specifics of House contests; thus, the House seems more promising. They cannot name the seats the GOP would have to lose to lose the House. If they could, they would find themselves naming many members most think are secure. A switch of the House still seems plausible, in other words, only because details are lacking.

History indicates that when the House switches, the Senate switches, too. Our knowledge of congressional elections implies that this is not coincidence. Accordingly, we can conclude that the safety of the GOP Senate strongly implies the safety of the GOP House. Further, we can issue a challenge to pundits who think the Democrats will take the House. They have an additional burden of proof: they must either indicate that the Senate will switch or why 2006 will be the first exception to a 92-year rule.

© 2000-2006 All Rights Reserved

Email Friend | Print | RSS | Add to | Add to Digg
Sponsored Links

Jay Cost
Author Archive