Sabato's Crystal Ball

Big and Little Nothings

Reviewing this cycle's mostly minor but occasionally major mistakes so far

Kyle Kondik, Managing Editor, Sabato's Crystal Ball April 24th, 2014

Yes, we know reporters have to react to news and find ways to make it relevant, but pardon us if we didn’t gag a little bit seeing headlines about the potential impact of Chelsea Clinton’s pregnancy on her mother’s potential presidential campaign. Some said the baby was timed for the campaign — because everyone knows a grandkid on the knee is a guaranteed vote-getter. (That’s why Mitt Romney won in a 2012 landslide.) Others suggested the opposite: Hillary Clinton was all ready to run until this news broke: Now she and Bill will want to babysit instead of barnstorming in Iowa (puh-leeze).

The minor media blip got us thinking about some campaign news from 2014: negative stories or gaffes that have at times popped up about this candidate or the other. Do these developments matter?

In most cases, no.

Clearly, some things that happen during campaigns change the game. Impolitic remarks on social issues by Todd Akin (R-MO) and Richard Mourdock (R-IN) probably cost them Senate seats in 2012. It’s harder to prove, but ethics problems might have fatally harmed Shelley Berkley (D-NV) in her narrow loss the same year. In all three cases, a more generic and less flawed candidate would probably have won.

Actual developments and/or mistakes will take their toll on candidates this cycle. Some already have, while others definitely didn’t. We separated these incidents into three categories: Ones that don’t matter, ones that might matter, and ones that do matter.

After checking out our list, e-mail us at to let us know which ones we might have missed. If we get some good responses, we’ll run an update next week.

Three that didn’t matter

1. Alison Lundergan Grimes’ campaign rollout: Remember back in the summer of 2013, when Kentucky Secretary of State Alison Lundergan Grimes (D) messed up her campaign rollout and, perish the thought, didn’t even immediately have a campaign website? No? That’s OK, because no one else remembers either. In fact, it seems like nothing that has happened throughout the first nine months of the expected clash between Grimes and Senate Minority Leader Mitch McConnell (R) has actually moved the needle. Just check out HuffPost Pollster’s tracking of the relatively tiny movement in the race:

Ultimately, we think McConnell has a significantly clearer path to a plurality than Grimes in this conservative state despite his unpopularity. Many Democrats differ. We’ll see — but any event that has or will have an actual impact on the race’s trajectory, outside of Grimes’ long-ago decision to actually run, probably hasn’t happened yet.

2. McConnell’s minor missteps: Speaking of McConnell, his campaign operation has had its share of minor headaches. His campaign manager was caught on tape complaining about the job. “McConnelling” became a new meme thanks to some stock footage the campaign put on the Internet for use by third-party groups. And the minority leader evoked memories of Michael Dukakis by awkwardly clutching a rifle at the Conservative Political Action Conference, among other things. But none of these stories have apparently meant much in McConnell’s primary race against Tea Partier Matt Bevin, who appears to have little chance to upset the five-term incumbent. Bevin’s own troubles, particularly his seeming support for the 2008 bank bailout that McConnell’s forces gleefully spread far and wide, seem more significant. So is the fact that it’s just really hard to beat an incumbent senator in a primary.

3. FitzGerald botches running mate: There’s not much evidence that a presidential running mate impacts the election to any major degree, so it stands to reason that a lieutenant gubernatorial nominee is even less important to voters. Presumptive Ohio Democratic gubernatorial nominee Ed FitzGerald (D) had to dump his initial running mate, state Sen. Eric Kearney (D), in December after Kearney’s tax problems became apparent. FitzGerald picked Sharen Neuhardt, a former congressional candidate and abortion rights advocate, to replace Kearney.

If you believe Democratic pollster Public Policy Polling, FitzGerald is tied with incumbent Gov. John Kasich (R) and has been for months. If you believe nonpartisan Quinnipiac, as well as Ohio’s voting history and the opinions of many people who know the state well, Kasich is leading. That was true six months ago — before the running mate switcheroo — and it’s true today.

Three that may or may not matter

1. “Entitlement” and Mark Pryor: The embattled Arkansas Democratic senator made a seemingly notable gaffe in early March when he said this of his opponent, Rep. Tom Cotton (R, AR-4):

“I think that is part of this sense of entitlement that he gives off. It’s almost like, ‘I served my country, let me into the Senate.’ That’s not how it works in Arkansas.”

Since the comment, though, the news for Pryor has been almost uniformly good. Five straight polls released this month — two independent surveys, two affiliated with Democrats or liberal causes and one Republican poll — showed Pryor either leading or tied with Cotton, and there’s been a notable uptick in opinion about his fortunes. The Republican poll, conducted by Harper Polling for Karl Rove’s American Crossroads, shows a clear trend: That poll had Pryor down six in late January but tied earlier this month. It seems fair to say that the gaffe didn’t hurt his numbers. But Cotton has a decent new ad out referring to the Pryor comment and his own military service, so perhaps it will remain in the bloodstream.

2. Bruce Braley’s trouble on the farm: Rep. Bruce Braley (D, IA-1) stepped in it when he insulted Sen. Chuck Grassley (R-IA) and farmers in general in a speech to lawyers. But how, exactly, did that change the race? Braley’s fate remains heavily tied to the Republican nomination process, which will go to an activist-dominated convention if no one gets over 35% in a June 3 primary, and to the overall political environment, which could hurt him in a swingy state that only quite narrowly leans Democratic. We knew all that after Sen. Tom Harkin (D) retired and no big names stepped up on the Republican side. Suffolk University polled the race in early April after the story broke and found results very similar to previous surveys from other pollsters: Braley was in the high 30s, and his likeliest and strongest potential opponents — businessman Mark Jacobs (R) and state Sen. Joni Ernst (R) — were around 30%.

If Braley had said this in late September as opposed to late March, when he could be facing a single, flush-with-cash opponent and a more tuned-in electorate, it would likely be a bigger problem. That said, a top Iowa observer insists to us that it’s important, and a political scientist who typically thinks such mistakes are overrated thought at the time that Braley had possibly done himself in. We don’t doubt that Republicans will tie this bell around the Braley cow all the way to November, and he’s not going to like it. The jury’s out on this one, but Braley was never as secure as some thought before the gaffe or as endangered as some thought after it.

3. David Perdue goes back to school: The Georgia Republican Senate candidate bashed fellow candidate Karen Handel by indirectly mentioning that she doesn’t have a college degree. (He belatedly apologized.) It’s possible that Handel, who hasn’t raised much money and who, it’s probably fair to say, has been a disappointment to national Republicans, will benefit from the much-needed free media the kerfuffle gave her. If Handel makes the likely primary runoff, and either Perdue is her opponent or he misses the runoff altogether (he’s leading in polls at the moment), then maybe this was a big moment. Otherwise, probably not.

Three that do matter

1. Milton Wolf’s X-ray commentary: Two candidates competing for the Republican Senate nomination in Kansas, which has been tantamount to election since 1938, have both suffered through a major controversy. Incumbent Sen. Pat Roberts (R) played right into Wolf’s hands when he admitted that he didn’t have a home in Kansas. This set off alarm bells with observers because residency issues likely played a big role in Sen. Richard Lugar’s (R-IN) primary loss in 2012. But then Wolf’s problems overshadowed Roberts: The insurgent, a physician, had posted morbid photos of X-rays on Facebook and made snarky remarks about them. Ultimately, the latter matters more: Wolf’s only chance to win was to run a flawless race against Roberts, and the X-ray disaster zaps any problems Roberts might have.

2. The kissing congressman: Rep. Vance McAllister (R, LA-5), who was taped kissing an aide after his recent special election to the House, is probably on his way out one way or the other. He has raised almost no money recently, and state and national GOP leaders want him to resign. If McAllister runs for reelection, he would be an underdog in Louisiana’s all-party primary in November, and even if he made it to a December runoff, it’s hard to see how he wins.

3. Doug Gansler’s string of mistakes: To be clear, Maryland Attorney General Doug Gansler (D) was always an underdog against Lt. Gov. Anthony Brown (D) in the Democratic primary to replace outgoing Gov. Martin O’Malley (D). But Gansler didn’t do himself any favors with continual mistakes, like appearing at a reportedly boozy teenage house party and, just recently, taking a shot at Brown’s military service that’s not all that dissimilar to what Pryor said about Cotton in Arkansas. The difference is that Gansler was and is very much behind (unlike Pryor) and it’s just the latest of many problems. Like Wolf in Kansas, Gansler needed to run a near-perfect race to win the nomination against Brown, whom O’Malley supports and who has consistently led polls. He has failed to do so.


The overall point is that while some of these gaffes and errors may matter, most won’t. Really, the campaign mistakes we’ve identified as mattering the most are mainly the ones in primaries. Wolf and Gansler are running underdog nomination bids, and McAllister is running in a heavily Republican district. If pushed, residents in LA-5 might pick McAllister over a Democrat, just like voters did when they sent damaged Rep. Mark Sanford (R, SC-1) to Congress in a special election last year. In 2012, Rep. Scott DesJarlais (R, TN-4) defeated a Democrat even after news broke that the doctor had pressured his one-time mistress to get an abortion. DesJarlais is an underdog in his primary this time.

As we saw most notably in the 2012 Republican presidential contest, primary voters shift more easily because they are picking among members of their own party with whom, presumably, they agree on most of the issues. General election electorates are more consistent, as the remarkably stable state-level polling in the 2012 Obama-Romney contest demonstrated.

Political observers always debate whether voters are, pardon our bluntness, stupid. If they were, they would be more likely to let many of the developments described above affect their voting. While cynics, particularly conservative ones, argue that the parties govern in similar, budget-busting ways, Democratic and Republican candidates are mostly and diametrically opposed on all sorts of issues. A voter who changed her mind because the candidate she otherwise supported made an occasional boneheaded comment or whose announcement press conference didn’t go all that well would be, in many cases, basing a vital, stark choice on fairly trivial matters.

We know that voters actually don’t gravitate between the parties all that much: As UCLA’s Lynn Vavreck recently and smartly noted, voters generally don’t switch from one party to the other in different elections. Sure, that’s proof that American voters are partisan, but it’s also evidence that they aren’t as affected by gaffes and “game changers” as campaigns and the press think.

How Veterans Vote

Not that differently from the rest of the electorate

Geoffrey Skelley, Associate Editor, Sabato's Crystal Ball April 24th, 2014

In Arkansas’ Senate contest, Rep. Tom Cotton (R) has a new ad that goes after Sen. Mark Pryor (D) for his comment that Cotton feels a “sense of entitlement” because of his military service. Cotton humorously utilizes his Army drill sergeant to talk about how his time in the military taught him “accountability, humility and putting the unit before yourself,” all qualities that voters might desire in a senator.

Cotton’s military background and Pryor’s ill-considered remark about it could conceivably matter more in Arkansas than in some other places (though there’s little evidence it has mattered so far). Based on 2013 estimates of adult and veteran populations, the Razorback State has the 13th-largest percentage of veterans as a proportion of its adult population (11.1%). That means there are a fair number of potential veteran voters in Arkansas: about 250,000, although the actual number who will turn out in a midterm will be far less. Hypothetically, these voters might be more receptive to Cotton’s candidacy because of his military background.

Examining Arkansas’ veteran population naturally led us to look at the country as a whole. Below is a map of states colored by what portion of their adult populations are veterans.

Map 1: Veterans as a percentage of adults by state, 2013 estimates

Note: Doesn’t include adult and veteran populations of Puerto Rico and other U.S. territories. Veteran data does not include active duty members of the armed forces. For a PDF of state data, click here.

Sources: U.S. Census Bureau for 2013 adult population estimates, U.S. Dept. of Veterans Affairs for 2013 veteran population estimates.

Alaska has the highest percentage of veterans in the country, followed by Virginia. The South and Interior West have higher percentages of veterans as a portion of the population compared to much of the rest of the country. But it’s important to remember that some states, such as California, Texas and New Jersey, have larger non-citizen populations, which impacts the proportions to some degree. Although some non-citizens decide to serve in the military as a means of gaining citizenship, most do not.

Of course, just because Cotton served in the military doesn’t mean that veterans are more likely to flock to him. As National Journal discussed in the lead up to the 2012 election, veteran voting behavior is more heavily influenced by factors such as party affiliation, ideology and religion, just like voters with no military background. A look back at past presidential results shows this: Veteran voters narrowly backed Bill Clinton in 1992 over George H.W. Bush (albeit in a three-way race), a World War II hero, and then in 2004 supported George W. Bush by 16 points over John Kerry, a decorated Vietnam veteran. In 2008, veteran John McCain did win veterans by 10 points, but as we’ll see, this was mainly a function of party identification and ideology. Unfortunately, the 2012 national exit poll did not ask broadly about military service, only asking it in Virginia, where President Obama and Mitt Romney tied 49%-49% among the 15% of the Old Dominion electorate that had served previously or who served at the time. Regrettably, 2008 Virginia top-line data is unavailable for the same question, making a comparison impossible.

Nevertheless, it’s likely that Cotton’s Republican Party identification will aid him more among veteran voters than will his veteran status. In Arkansas, 84% of veterans are non-Hispanic white, and white voters, particularly those in the South, are more likely to vote Republican. Moreover, the overwhelming majority of veterans are male, and men are more likely to vote Republican as well. Nationally, a 2011 Pew Research Center study found that nearly half of all veterans (48%) say they are politically conservative, compared to 37% of the public in the same poll. Veterans are generally more likely to identify as Republicans than Democrats, which again can perhaps be tied back to race. As Table 1 shows below, veterans are whiter than the electorate as a whole, and therefore it makes sense that they might be more inclined to vote Republican.

Table 1: National demographic makeup of 2012 electorate compared to veteran population

Sources: CNN, U.S. Dept. of Veterans Affairs

But like other aspects of American life, the demographic makeup of our veterans will also change as the nation becomes more diverse. In fact, the U.S. Department of Veterans Affairs projects that the veteran population will be 66% non-Hispanic white, 11% Latino, 16% African American and 2% Asian by 2040. That will still be whiter than the electorate as a whole if trends continue, but the veteran population will be more diverse than it is today.

Exiting the House

The many paths to ending a career in Congress' lower chamber

Geoffrey Skelley, Associate Editor, Sabato's Crystal Ball April 17th, 2014

Over the past 40 years, there have been many ways to leave the U.S. House of Representatives. Specifically, nine different methods. The main ones, beyond losing a primary or general election, are to retire or run for another office. But a member can also do one of the following: be appointed to another office, resign, be expelled, pass away or, in the rarest of instances, have the House vacate one’s seat.

So far, 50 members of the 113th Congress have either left office or signaled their intentions to leave at the end of this cycle. The manner in which they have left or plan to leave the House varies. Two already found paths to the U.S. Senate: Then-Rep. Tim Scott (R-SC) was appointed to the upper chamber and then-Rep. Ed Markey (D-MA) won a special election to replace Secretary of State John Kerry. Another 17 are in the midst of running for other offices that will preclude them from running for the House again — 13 are running for the Senate (or ran, in Republican Rep. Steve Stockman’s case), two are running for governor, one is seeking a lieutenant governorship and another is hoping to become a county supervisor. Most of those exiting the House (24) will do so by retiring at the end of the term, while six have already beat them to the punch by resigning. Lastly, the late Rep. Bill Young (R) died, necessitating the hotly-contested special election in March to fill his Florida seat.

Despite all that, the degree of turnover in the House this cycle is not unusually high. Over the last 40 years, an average of 70.4 members has exited the House for one reason or another each two-year cycle. That’s about one-sixth of the total House membership every cycle. At 50 exits so far, this Congress still has a ways to go in order to produce even an average level of turnover. Of course, there might still be additional retirements or resignations, and some incumbents will lose primary and general election contests. However, while this cycle’s total will go up, it remains to be seen whether or not it will reach or surpass the average number of departures.

Table 1 below lays out data from the 93rd Congress in 1973-1974 to the present. (Thanks to Roll Call for producing its invaluable “Casualty List” over this time period.)

Table 1: Exiting the U.S. House, 93rd Congress to 113th Congress

Sources: Roll Call “Casualty List,” Crystal Ball research

Notes: *These numbers are very likely to change as some running for other offices will be defeated and some members may lose in primary or general elections; averages for these columns do not include the totals for the 113th Congress. **For the 113th Congress, the figure is the number who are running or have run for other offices (Republican Rep. Steve Stockman recently lost in a Senate primary); for all others, it’s the number defeated in runs for other offices. The data above excludes representatives of U.S. territories. To access a PDF version of this chart, click here.

In terms of retirements, the 113th Congress has just ticked above the mean, with 24 retirements versus an average of 22.7 over the past four decades. Similarly, the six resignations are also above average. Meanwhile, the single appointment to another office (Scott) about matches the 40-year mean, and the number of deaths is well below average. As for the number elected or defeated in bids for other offices, those figures will remain incomplete for the time being because we don’t know yet which soon-to-be-former members of the House will win or lose. So far, 41 members are retiring at the end of this term or running for other offices with general elections this November, which is slightly ahead of the average of 36.9 over the last 40 years.

There are plenty of eye-catching data points in this table. On average, more representatives lose races for other offices than win (8.6 versus 6.0). This may be partially because members sometimes find themselves facing off against other members in their party primaries or in general elections. This can happen not only in member-versus-member primaries just after redistricting but also in bids for other offices. Just think of the open Senate race in Georgia this cycle, where three GOP representatives are running for their party’s nomination. Although one may add on to the “elected to other office” total, at least two and possibly all three will wind up in the “defeated for other office” category when all is said and done.

For all the talk of how awful being on Capitol Hill is, this Congress is a long way from matching the 102nd (1991-1992) for members who either retired or ran for other offices. That term saw 52 members hang up their lapel pins and ride off into the sunset, while another 13 ran for offices with elections in November 1992 (that total doesn’t include Republican Rep. Steve Bartlett, who won the Dallas mayoralty in November 1991). The total number of retirements and runs for office elected in November that cycle, 65, is more than one-and-a-half times 41, which is the current total in this Congress. To put that sum in perspective, 65 retirements and standard runs for other office is a larger figure than the number of net seats House Democrats lost in the 2010 Republican wave (63). The 102nd Congress is also fascinating because it featured the largest number of total exits (116) of any in the past 40 years despite 1992 not being a very large “wave” election year in the manner of 1974, 1994, 2006 or 2010. One explanation is that 1991-92 was a redistricting cycle, which is a once-in-a-decade wrinkle that can contribute to House turnover — three of the four redistricting cycles listed above (1981-82, 1991-92 and 2011-12) featured above-average House turnover. Another explanation might be the infamous “House banking scandal,” in which many members were found to be, in essence, kiting checks without penalty in the House bank.

That said, and while its conclusion is far from written, at this point it appears the 113th Congress’ non-House election exit total (i.e. those leaving for reasons other than losing a House primary or general election) may exceed 50 (following Republican Rep. Tom Petri’s retirement announcement last week, the total is exactly 50 right now). That sum is higher than the average of 45.6, the third straight Congress to be above average. So perhaps there is a bit of truth to the “Congress is a miserable place to be” line after all. To bolster that argument, one could also note the large number of resignations over the last three congresses — a total of 26 — as opposed to the average of four per Congress over the last four decades.

But at least no one in this Congress has been expelled. Two members in the past 40 years, Rep. Michael Myers (D-PA) in 1980 and Rep. Jim Traficant (D-OH) in 2002, have been kicked out of the House. Myers was the first representative since the Civil War to be expelled after his involvement in the Abscam scandal; his acceptance of bribes in exchange for influence eventually led to a federal conviction and jail time. Two decades later, Traficant was convicted of 10 felony counts, including bribery, tax evasion and accepting kickbacks. He went on to serve seven years in prison, which didn’t stop him from running a quixotic independent campaign for his old House seat. Of course, others may have joined them in this category if they hadn’t opted to leave on their own accord in light of their transgressions (i.e. resign).

Lastly, it’s worthwhile to point out the oddest number in the table: the single instance where a seat was “vacated,” which occurred during the 97th Congress. Rep. Gladys Spellman (D-MD) suffered a heart attack just days before the 1980 election was to take place and fell into a coma. She was popular, and her popularity was enhanced by a sympathy vote: On Election Day she was reelected with over 80% of the vote despite being comatose. Unfortunately, it became clear that she would not awaken any time soon (and she never did, sadly). This situation prompted the House to declare the seat vacant, the only time medical circumstances have ever led the chamber to do so. Steny Hoyer (D) defeated Spellman’s husband in the Democratic primary and won the special election. (The power of widows to win the House seats of their departed husbands has long been noted, but this apparently didn’t extend to a widower.) Hoyer is now the House Democratic Whip, having served as majority leader when Democrats controlled the chamber during Nancy Pelosi’s speakership.

It just goes to show that while there are a lot of odd ways to exit the House, there are also some odd ways to enter it.