The Rule of 13 in House and Senate Elections


If you want to predict House or Senate elections, a useful notion is what I call the Rule of 13. It says that if a district is misaligned with your partisanship by more than 13 points, then, to a close approximation, you have zero chance of winning that district. The rule predicts the following: (i) Mark Pryor is sure to lose his Senate reelection bid in Arkansas, (ii) Mitch McConnell is sure to win his reelection bid in Kentucky, (iii) if voters become convinced that challenger Greg Orman is, for all intents and purposes, a Democrat, then Pat Roberts is sure to win his reelection; (iv) although Alaska, Georgia, Louisiana, and North Carolina are conservative states, they are not conservative enough to invoke the Rule of 13; accordingly the Democratic candidates in those states at least have a chance of winning; (v) although the West Virginia 2nd and 3rd House races are called “tossups” by some prognosticators, the Rule of 13 says that the Republican candidates (Alex Mooney and Evan Jenkins) will win for certain.

The Rule of 13 is formally defined as follows. First, define the partisan index of a district according the most recent presidential vote in that district. For example, consider the situation of Rep. Mike Ross (D-Ark.) near the end of his sixth term in office, 2010-12. At that time, the most recent presidential election was the 2008 race between John McCain and Barack Obama. In Ross’s district (based on lines redrawn after the 2010 census) McCain received 166,247 votes and Obama received 103,478 votes. McCain’s share of the two-party vote in the district was thus 61.6%. Meanwhile, McCain’s two-party vote share in the nation was 46.0%. Define the partisan index of Ross’s district as the difference of those two numbers. Thus, the district’s partisan index was “Republican 15.6.” (The Cook Political Report constructs a similar “Partisan Voting Index,” except it bases its number on an average of the prior two presidential elections. Some research I’ve conducted suggests that the partisanship of a district follows a random walk, which implies that only the most recent presidential election is relevant in predicting the political views of a district; prior elections do not provide any more information.)

The Rule of 13 is the following: “If you’re a Democrat and your district’s partisan index is Republican 13.0 or higher or if you’re a Republican and your district’s partisan index is Democrat 13.0 or higher, then you have no chance of winning.”

The rule thus would predict that Ross would have no chance at winning reelection in 2012. Perhaps following the rule, Ross declined to run for reelection.

In 2012, two other House members were misaligned with their districts by at least 13 points—Heath Shuler (D-N.C.) and Dan Boren (D-Okla.). Like Mike Ross, Shuler and Boren chose not to run for reelection. Like the House, the 2012 Senate elections saw no violations of the Rule of 13.

In 2012, two House incumbents were close to being misaligned by 13 points or more. One was Nick Rahall (D-W.V.), whose district had a partisanship index of “Republican 12.5.” Despite the large misalignment (yet not quite meeting the criterion for the Rule of 13), Rahall won reelection by 54-46 (albeit against an opponent he’d defeated by 32 points in 2004). The other House member was Jim Matheson (D-Utah), whose district had a partisan index of “Republican 11.5.” He defeated Mia Love, 48.8% to 48.5%.

In 2012, in Rahall’s and Matheson’s districts, Romney performed much better than McCain did in 2008. Consequently, the partisan indexes of those districts rose significantly—to “Republican 18.5” in Rahall’s district and “Republican 21.0” in Matheson’s district. (The latter number is a bit artificial since Mitt Romney is somewhat of a native son to Utah. However, Romney’s vote share, compared to McCain’s, rose by about one and a half percentage points more in Matheson’s district than it did in the rest of the state. Thus, it appears that Matheson’s district genuinely became more conservative than it was in 2008, and it appears that, even without the native-son effect of Romney, the district would be above or at least very near the Rule-of-13 standard.)

Matheson has chosen not to run for reelection in 2014, and it appears that Mia Love will win the seat easily. Rahall, who represents West Virginia’s 3rd district, is in a tough fight against Republican Evan Jenkins. Although RealClearPolitics calls the race a “tossup,” the Rule of 13 says that Jenkins will win for certain.

The 2nd district of West Virginia is another interesting illustration of the Rule of 13. The entire state of West Virginia has been trending rightward. For instance, eight years years ago its partisan index was “Republican 5.2”; four years ago it was “Republican 10.4”; and now it is “Republican 15.7”. Perhaps understanding the Rule of 13, the current representative of the 2nd district, Shelley Moore Capito, decided to challenge Jay Rockefeller for his Senate seat. Rockefeller might also have understood the Rule of 13—soon after Capito’s announcement, he decided not to run for reelection.

That left an opening for Capito’s House seat, which is currently being sought by Republican Alex Mooney, a former Maryland state senator, and Nick Casey, a former state chairman of the West Virginia Democratic Party. The partisan index of the district is “Republican 13.2”. Thus, the Rule of 13 says that Mooney will win for certain. Nevertheless, Casey has mounted a vigorous campaign to paint himself as a moderate—he calls himself “pro-life, pro-gun, and pro-coal”—and RealClearPolitics calls the race a tossup. Still, my bet is with the Rule of 13, that the race really is not a tossup.

Sen. Mark Pryor is currently feeling the effects of the Rule of 13. When he won his first Senate race, in 2002, the partisan index of Arkansas was “Republican 3.1.” When he ran for re-election in 2008 the index was “Republican 3.7.” In both years Pryor won easily. Since the 2012 election, however, the index has become “Republican 14.2.” Over the last several days, many people have begun to predict that Pryor will lose. (E.g., within the last day or two, RealClearPolitics moved the race from its “tossup” column to its “leans GOP” column.) The Rule of 13, however, predicted a defeat for Pryor almost two years ago.

I have no theoretical reasons for choosing 13 as the magic number for the rule. Rather, I derived the rule strictly through observation—by “soaking and poking” the data until I found a number that seemed to work. I believe the rule is novel—for instance, I have never heard any pundit, consultant, or scholar discuss any similar rule.

Related, I believe that the rule is far from obvious to most political observers. At least, the rule was not obvious to me several months ago, before I began trying to construct it. For instance, in 2010, I was very impressed with Joel Pollak when he decided to run for Congress. Although his opponent, Jan Schakowsky was an entrenched incumbent in a very Democratic district, I thought that Pollack’s superior intelligence and charisma would overcome the latter factors. However, the partisan index of the district was “Democratic 19.6.” If I had known about the Rule of 13, I could have foreseen that no level of intelligence and charisma would allow Pollak to win. A similar instance occurred the same year, when Republican Star Parker challenged Rep. Laura Richardson for Congress. Here again, I thought Parker’s outstanding talent gave her a chance to win, and I even donated to her campaign. However, the partisan index of Richardson’s district was “Democrat 27.3.” The Rule of 13 predicted that Parker had no chance, and it was correct. She lost 68% to 23%.

Although rare, there are some exceptions to the Rule of 13. For instance, Arkansas’s Mike Ross won in 2010 even though the partisan index of his district was “Republican 13.9.” Similarly, Oklahoma’s Dan Boren won in 2010 even though partisan index of his district was “Republican 19.1.”

Further, the rule does not work quite as well for governors’ elections. For instance, Arkansas Gov. Mike Beebe won reelection in 2010 even though the partisan index for his state was “Republican 13.9.” In 2002, Mitt Romney won governor of Massachusetts even though the state’s partisan index was “Democratic 14.5.” For governors’ races a “Rule of 14” or a “Rule of 15” seems to be more accurate.

Still, by and large, there are few exceptions to the Rule of 13. It, accordingly, is a useful tool for predicting House and Senate elections. It is also useful as a prescriptive measure. If, say, any young Ricocheters are thinking about running for Congress someday, here is some advice: First, start thinking about a district now. Second, make sure that the district is no more than 13 points out of line with your partisanship.

There are 5 comments.

Become a member to join the conversation. Or sign in if you're already a member.
  1. user_82762 Thatcher

    Dr. Groseclose,

    First, thank Gd we have you on our side.  The information is very precious and we all should listen and try to apply it with vigor.

    However, I guess I just can’t help but think that there is some other ‘force’ at work. Something unseen and unmeasurable.  Perhaps when your papers have passed peer review and you have the time, you feel it too.



    • #1
  2. MJBubba Member

    The moral of the story is to work like hell to elect Republicans to State Senate seats where they can control the gerrymandering re-districting.

    • #2
  3. Palaeologus Member

    This is fascinating stuff, Tim.

    I’ll note that both of your listed “Rule of 13” congressional outliers were multi-term incumbents.

    Not only did they survive in opposition dominated districts; these Dems survived in a GOP wave cycle.

    I don’t know squat about either candidacy beyond that. Perhaps their respective opponents were amazingly incompetent… or perhaps incumbency is weirdly under-appreciated by electoral analysts.

    • #3
  4. user_252181 Moderator

    So in Oregon Mitt got roughly 44% of the vote, vs nationwide 46%, making Oregon D +2.  Than means Wehby has a chance?  It doesn’t look like it.

    • #4
  5. Owen Findy Member
    Owen Findy

    What in God’s name is the point of predicting?  (Fun is a perfectly legitimate purpose, BTW, but I just don’t enjoy thinking or reading about what might be, when so much hangs in the balance.)

    What will be, will soon enough be, and then we’ll know.

    I also am convinced that most economic predictions are nearly worthless and it ticks me off that so much breath is spent on it.  (I’m only hedging my qualifiers, here, because I know my own knowledge is imperfect….)

    • #5

Comments are closed because this post is more than six months old. Please write a new post if you would like to continue this conversation.