Tag Archives: Specter

Revisiting Earlier “Specter-lations” on the Eve of the Pennsylvania Senate Primary

One advantage political scientists have over the media or pundits is that we have the luxury of revisiting earlier projections/explanations in light of new data or theorizing.  This is an advantage when writing a blog that tries to focus on the fundamentals driving political events rather than simply opining on the latest topical event.  In particular, when my projections prove wrong (a not infrequent occurrence as long time readers know), I see that as an opportunity to learn in the hope of increasing our understanding of political processes and outcomes.    At least that’s the goal.

In this vein, with voters scheduled to go to the polls in tomorrow’s crucial Democrat Senate primary in Pennsylvania (and in Senate primaries in three other states) , I thought it worthwhile to revisit several posts I wrote almost two years ago regarding Senator Arlen Specter’s decision to switch from the Republican to the Democrat party.  At the time I suggested that Specter’s switch was driven by electoral considerations, not ideological ones.  He was reacting to polling data indicating he would lose to the more conservative Pat Toomey in the Republican primary, so he changed venues to run in the presumably less challenging Democrat primary.  That switch, I suggested, would not appreciably change his Senate voting habits.  Here’s what I wrote:

“If I’m right, Specter’s voting record should keep him somewhere in the middle of the Senate pack, ideologically, over the course of the next two years.  If my poli sci colleagues are correct, he should move sharply left closer to the middle of the Democratic Senate voting ranks (say, into Dianne Feinstein/Harry Reid territory).   So we can revisit this issue in two years!”

Almost two years have passed.  How well did my prediction turn out?  Not very well at all.  In fact, based on Jeff Lewis’ roll call data, summarized in this chart by Simon Jackman, Specter’s voting record eventually shifted Left after his party switch, toward the middle of the Democrat Senate caucus, very close to Feinstein territory.  (Yes, this chart summarizing ideological placement based on roll call votes does not differentiate between significant and less significant votes, but it is a widely-used marker of relative ideological placement nonetheless.)


Why did I suggest Specter’s voting record would not change very much, and why was I wrong? When in the days after his switch Specter was one of four Democrats to vote against the Obama-backed budget resolution, thus appearing to validate my initial projection, I wrote: “So far, his biggest fear is not from the unknown candidate in the Democratic primary (he’s evidently counting on the Democratic Party clearing the field for him) – it’s from his likely Republican opposition in the general election.”

In my defense, then, at the time I made my initial projection that Specter would not change his voting spots, I assumed he was primarily focused on a challenge from the Right, not the Left.  As I warned at the time:  “Of course, if he gets a credible challenger on the Left in Pennsylvania, then all bets are off!”  However, I discounted a primary challenge in the belief that the Obama administration had negotiated a deal with Pennsylvania Democrats to give Specter an open field to the Democrat nomination.

That assumption turned out not to be true.  Since Specter’s switch, we have been reminded just how little clout presidents have in local party affairs, and  how the decision to run on the party label for a Senate seat in American politics is no longer controlled by party leaders.  Specter in fact faces a very credible challenger on the Left in the person of Joe Sestak. At the same time, my projection was made before the full scope of the anti-incumbent wave that now appears building in the run up to 2010 became evident. Not only does Specter face a primary challenge – if the latest polls are to be believed, he is in danger of losing! Although Obama has kept his promise to back Specter (and has even run ads on Specter’s behalf), the latest Pollster.com composite polling graph has the Democrat Senate primary in a virtual dead heat between Specter and Sestak.

Now, there are between 10-15% of undecided voters in these polls, so Specter may yet pull this primary out. Nonetheless, this is a far cry from the “open Democrat field” on which I premised my earlier projection. The irony is that Specter’s move left did nothing to dissuade Sestak from opposing him, and yet has likely made it more difficult for Specter to win the general election; even if he beats Sestak in the primary, polls indicate he will lose the general election to the Republican Pat Toomey.  It would have been far better for Specter’s electoral fortunes, I think, had he maintained his traditionally moderate voting record.

There are two important points to take home here. First, Specter’s potential demise – either at the hands of a Liberal Democrat or a Conservative Republican – is another reminder why we have a polarized Senate.  It’s not because voters in a general Senate election are more polarized – it’s because they often are forced to choose between two relatively polarizing candidates.  One could argue that neither Sestak nor Toomey are as broadly representative of the median voter in Pennsylvania as is Specter, but that may not matter if Specter’s not on the ballot in the general election.

Second, the signs continue to suggest that the anti-incumbent sentiment is shared by enough voters – not all of them Tea Party supporters – to make 2010 a potential “wave” election year.  Of course, as I have continually cautioned, it is easy to make too much of individual races. If Specter loses, pundits will undoubtedly suggest this is another indication of Obama’s weak political influence. Given the general anti-incumbent trend, and the historical inability of presidents dating back to FDR to influence midterm races, I’m not sure it’s accurate to lay the blame at the White House’s doorstep.

Nonetheless, after tomorrow’s Senate primaries (along with Pennsylvania, there are primaries in Kentucky, Arkansas and Oregon), and the Special Election to fill the U.S. House  seat in Pennsylvania’s 12th district formerly held by Representative John Murtha, Jr. (Murtha died last February), we will have a bit more information with which to try to assess the lay of the electoral landscape.  If I can, I’ll try to summarize the results when the polls close Tuesday night.

Meanwhile, with only four predictions in, I’ve extended the window on forecasting the Kagan confirmation vote for another week.  Remember, if you are a graduating senior, your job prospects will undoubtedly be helped if you nail this projection.  I am told that Mike Norris, a previous contest winner, used his victory to get a lucrative (six figures) job working for Bill Gates as a technology forecaster.

Does the “Specter” of a filibuster-proof Senate really help the Democrats?

In an earlier blog comment, Jack Goodman asked me about a Arlen Specter-Tom Ridge matchup in the 2010.  I suggested that Ridge, a former Governor in that state, was the type of moderate Republican who could defeat Specter, which was one reason why I didn’t think Specter’s Senate voting would move very far Left, ideologically, despite his new party label.  I have no idea if Ridge is running, but Quinnipiac just released polling data (see here) confirming my hunch.   In the first survey of Pennsylvania voters since Specter’s defection, it shows Specter and Ridge running neck-and-neck; Specter leads 46-43, with 8% undecided – a lead that is about the same as the poll’s 2.9% margin of error.

In contrast, the same poll shows Specter trouncing Republican congressman Pat Toomey, 52-33, with 10% undecided.

Against either Ridge or Toomey, Specter enjoys strong support among Democrats; the numbers are 84-5 against Toomey, and 78-14 over Ridge.  Among Republicans, however, it is the reverse; here Specter loses 74-14 to Toomey and 82-10 to Ridge.

The biggest difference between having Toomey as an opponent as opposed to Ridge – and why Ridge is such a formidable opponent – comes among independents.  They favor Ridge over Specter 47-37, with 11% undecided.

This is why you see so little change in Specter’s voting record to date despite the party switch; lacking a credible Democratic primary opponent, and with strong support among Democratic voters, he’s not very worried about opposition from the Left.  (The favorable/unfavorable split among Democratic voters toward Specter is 77-8!) It’s the moderate voters he’s worried about, particularly if his opponent is Ridge.  Among independents, Specter’s favorability ratio is 51-35 – but Ridge’s is 62-17 (18% undecided). And independents are evenly split, 44-44, regarding whether Specter deserves to be reelected in 2010.  Hence my prediction that Specter would remain firmly entrenched as a swing vote in the Senate.

If I’m Ridge (and Republicans more generally), the following is the part of the Quinnipiac poll that I would find most encouraging:  only 41% of those surveyed thought that a Democratically-controlled filibuster-proof Senate was a good thing – 49% said it was a bad thing (with 11% undecided). And by a margin of 52-44, Pennsylvanians agree with the following statement: “Some people say that losing a Republican in the Senate is dangerous because President Obama and the Democrats will now be able to steamroll over the Republicans. Do you agree or disagree?”  That, in my view, is the issue on which Republicans can win back this Senate seat: the “Specter” of a Democratic supermajority in the U.S. Senate.  Pennsylvanians – and Americans more generally – are uneasy when power is concentrated in a single party. And that’s why Specter’s switch might not be all that beneficial to Democrats – if his party label changes, but his voting patterns do not, then it provides Pennsylvanian voters concerned about a concentration of power with a pretext to elect a “true” Republican Senator in 2010 to limit Democratic power at the national level. From this perspective, Democrats would have been better off if Specter remained a Republican.  The problem, of course, is that Specter likely would not have won as a Republican.

Note that this is one poll, and the election is more than a year away. Ridge has not, to my knowledge, even expressed any interest in running.  More importantly, it did not ask voters about a Ridge-Toomey Republican matchup – it’s not immediately clear to me that Ridge can make it out of the Republican primary if he’s matched up with Toomey.  Nonetheless, the data provides support for my contention that Specter is more worried about his right flank than his left.  And it provides a glimmer of support for those predicting a Republican comeback in the 2010 midterms.

Say it ain’t so, Dr. Joe! (And More Specter fallout)

Leading the news today is Dr. Joe Biden’s medical advice, issued on the Today show, that people should avoid riding in airplanes, or taking the subway, or using other forms of mass transit.  Never mind that health officials don’t agree with Dr. Joe – he’s the vice president!  With one foot-in-mouth moment, Dr. Joe singlehandedly counteracts the impact of the $800 billion stimulus bill. We love this guy!  (So does the President, I am certain.)

Now, in response to your inquiries, more on the Specter decision. Several of you have asked why I discount the political science research showing that a switch in party labels usually leads to a switch in voting patterns.  In fact, based on that research, most political scientists that I know believe Specter’s voting pattern will now move quickly and significantly to the Left (see here, for example).

The reason why I don’t believe this is necessarily the case is because that research is based on House switchers, not Senate switchers. In addition to the differences in how the two chambers operate (the House is designed to empower the majority party as a voting bloc, the Senate operates on a more individualistic basis), House districts are almost always smaller, and more ideologically homogeneous.  Senators, on the other hand, represent typically larger and more ideologically diverse states.  As a result, House incumbents usually are less vulnerable to electoral defeat (their reelection rates hover in the mid 90% range), and find it easier to stake out a more partisan voting position congruent with the majority of voters in their district.  Not so for Senators, who typically face stronger challengers who are more effective at using their voting record against them, and who must appeal to a more diverse electorate.   That’s why I am skeptical that previous research on party switching based on the House is applicable to the Senate.

But this provides the opportunity for a natural experiment.  According to Simon Jackman (see chart below), Specter now ranks as the third most liberal Republican in the current (111th) Senate.  (These rankings are based on the Nominate scores – moving Left, along the negative numbers, means you are more liberal, right toward positive numbers is more conservative. The Senators are broken into two columns to save space.)

If I’m right, Specter’s voting record should keep him somewhere in the middle of the Senate pack, ideologically, over the course of the next two years.  If my poli sci colleagues are correct, he should move sharply left closer to the middle of the Democratic Senate voting ranks (say, into Dianne Feinstein/Harry Reid territory).   So we can revisit this issue in two years!

So, why did Specter switch parties?  Largely because of the fallout from the stimulus bill – another reminder why bipartisanship is so difficult in today’s polarized climate.  A March Quinnipiac poll of Pennsylvania voters found the following (see poll here): “Overall Pennsylvania voters have a 45 – 31 percent favorable opinion of Sen. Specter, but he gets a 47 – 29 percent unfavorable score from Republicans. The Republican gets a 60 – 16 percent thumbs up from Democrats and a 41 – 35 percent positive from independent voters.” Much of that disapproval is rooted in Specter’s support of the stimulus package. “Specter’s support of President Barack Obama’s Stimulus Package wins 87 – 6 percent support from Democrats and 56 – 38 percent support from independent voters, while Republican voters disapprove 70 – 25 percent.”  It is not a good sign when your own party’s constituents disown you. As a result, he was likely to lose if he ran in the Republican primary against the less well known Republican Congressman Pat Toomey.  The Quinnipiac poll found Specter trailing Toomey 41 – 27 percent in a Republican primary for the 2010 Senate race, with 28 percent undecided.  A more recent Rasmussen poll (see here) of likely Republican primary voters has Specter getting crushed by Toomey, 51%-30%. More generally, Specter’s approval ratings began nosing downward after the stimulus vote, although he still has more favorable than unfavorable support among all voters.

There are two lessons to draw from this.  First, if Specter has negotiated an understanding with leading Democrats to clear the field of potential challengers in the Democratic Pennsylvania Senate primary (as it appears to be the case), then he has a strong shot at winning the general election for another term as Pennsylvania Senator.  But that wouldn’t happen as a Republican; he wasn’t likely to make it out of the Republican primary.

My larger point is to reiterate an earlier observation: it is often said by pundits, particularly those on the Left, that the Republicans have become the party of “NO”, unalterably opposed to anything Obama proposes.  But the reality, as Specter discovered, is that many Republicans on Capitol Hill face a potential backlash among their core Republican voters in the primary if they appear to support programs that involve increased spending and a greater government role in the economy.  From this perspective, Republican opposition to much of Obama’s program is perfectly logical (and quite predictable).  As an aside, note that Obama doesn’t seem to grasp this, at least not based on his public statements.  Last night during his press conference, he once again reiterated his reminder to Republicans that, “we won.”   Of course, as I’ve pointed out in previous posts, from the perspective of individual members of Congress, this is not the case – Obama ran behind most of them, and of course didn’t win very many Republican districts at all.

Aside to Jack, Vijay and Marty:  I don’t have polling data on Tom Ridge versus Specter, but I think Ridge is precisely the kind of Republican who would make it difficult for Specter to win in the general election. However, (for reasons Vijay alludes to in his comment on my earlier post) Ridge would also have to get through the Republican primary – not a sure thing, although he doesn’t have the baggage of a stimulus vote to defend.  Also, I don’t think he has lived in Pennsylvania in some time, although given his history there that may not make a difference to voters.  For what it is worth, here’s Ridge’s statement on Specter’s defection:

“I’ve known Arlen Specter for many years. In no way does his departure from the Republican Party diminish his long record of service to his country and to the Commonwealth of Pennsylvania.

“As Arlen will understand, my support is with my Party and with the people of my home state, whom I believe can best be served in the Senate by the GOP.

“So let’s begin the discussion of ideas with a respectful contest as Pennsylvania and the nation continue to work collectively through these challenging times.”

Does that mean Ridge is jumping into the race?  I have no idea.  As for the “Club for Growth” and PAC’s more generally, as my colleague Bert Johnson’s research reminds us, the power of PACs to shape elections appears to have diminished a bit compared to the relative impact of campaign contributions from strongly partisan individuals.  I hope to post on this more extensively.  In any case, I wouldn’t suggest that Specter is the victim of an orchestrated party purge, so much as he fell prey to diminishing enthusiasm among likely Republican voters.  There is a fascinating story here on why politics is increasingly polarized and I hope to discuss it more fully in later posts.

Thank God for Arlen Specter

Senator Arlen Specter’s decision to caucus with the Democrats has, temporarily, pushed the “100 days” storyline into the background, and for that we should all be grateful. (For my sole media effort to correct the  misperception that 100 days matter, see my brief comments with a Canadian journalist here.)  In addition to the blow it inflicts on the Republican Party, at least symbolically, Specter’s defection is another reminder of how difficult it is for moderates to exist in a deeply polarized Senate.  I went back and calculated Specter’s rank, using a generally accepted measure of voting-based ideology (Poole-Rosenthal scores) in each Senate session in which he served dating back to his first in 1981-83 (the 97th Congress).  When it comes to voting, Specter consistently falls in the middle of the pack, with an overall average ranking across 15 congressional sessions as the 52nd most liberal Senator (standard deviation of 4.3).  (Note that I include the President’s rank score in the data used to calculate Specter’s standing).

When the Senate is under Democratic control, the overall mean voting score shifts left, which tends to make Specter more conservative relative to his colleagues; his average rank under Democratic control is 57. Conversely, under Republican control, he ranks slightly more liberal, typically landing right in the middle as the 50th most liberal Senator. On the whole, then, we see that he’s a consistent middle-of-the-road Senator.

So, is his voting record likely to significantly change now that he’s a Democrat?  At first glance, one might think yes.  Indeed, several pundits have suggested as much based in part on political science research on previous party switchers. This research suggests that a switch in party leads to a switch in voting patterns.  But caution is in order; most of this research finds significant changes only in the House.  In the Senate, because of the small sample size, detecting historical trends is much more difficult.

For comparison purposes, I looked up the rankings of four other Senators who switched party affiliation in the modern era to see if it affected their overall Senate ranking.

Senator Years (Year Switch) Pre-Switch Rank Post-Switch
Richard Shelby 1987-2008 (1995) 61st most liberal 77th most liberal
Ben Campbell 1993-2004 (1995) 51 56th most liberal
Jim Jeffords 1989-2002 (2001) 52 52
Joe Lieberman 1989-2008 (2007) 37 34.5

Keep in mind that both Shelby and Campbell switched from Democrat to Republican, while Lieberman ran as an independent after losing the Democratic primary in 2006 but did not join the Republican Party.  After years as a Republican, Jeffords caucused with the Democrats in 2001 but called himself an independent.  So these aren’t strictly comparable to the Specter case.  (The other significant difference, as I discuss below, is that Specter’s switch is motivated by electoral reasons more than ideological ones.)

So based on this limited sample, and subject to all the caveats associated with nominate scores and using ranking, as opposed to absolute voting data (and the different time periods being compared), it appears that Shelby’s voting rank changed significantly, but the other three demonstrate less movement. When we think about the issue a bit more, it’s not surprising that this admittedly limited sample reveals no clear trend in voting patterns after a change in party affiliation among Senators, in contrast to Representatives in the House.

Remember that the House operates under rules designed to empower the majority party, so in periods of high polarization, a switch in parties should tend to lead to a switch in voting (much as political science research has found.) However, this is not necessarily the case in the Senate, where voting is less beholden to party, and much more driven by individual ideology conditioned by constituency and reelection.  In other words, the Senate is not strictly comparable to a parliamentary system, where party leaders often have the means to punish those who cross party lines by, for example, removing them from party lists at election time.  In the U.S. Senate, senators are largely beholden to no one except voters.  Note that based on Poole-Rosenthal scores from the last (110th) Senate, Specter is the third most liberal Republican Senator, topped only by the two Maine Senators Collins and Snowe.  Not surprisingly, all three Republican Senators already voted with the Democrats to pass Obama’s stimulus bill, and there wasn’t a darn thing the Republican Senate leadership could do about it.

All this is a long way of saying that I don’t expect Specter’s rather moderate voting patterns to change very much as a result of the change in party label, since he’s already signaled his willingness to cross parties and vote with the Democrats during the current session if it serves his electoral needs.  Indeed that is what this switch is about – Specter’s 2010 reelection bid.  By switching labels, he can run in the Democratic primary where he (so far) has at least a chance of reelection.  If he stayed a Republican, he was going to get trounced in the primary (the most recent poll had him down 21% among likely Republican voters to the more conservative Pat Toomey).  For Specter, becoming a Democrat was a matter of political survival – not a change in ideology. By negotiating a deal with the Democrats (including Obama’s implicit promise to back him in the Senate race), he’s hoping to avoid a challenge from the Left in the Democratic primary in two years.  If so, I expect him to continue to position himself as a swing voter in a state that is trending Democratic, with his vote determined by his calculations on its likely impact on his Senate reelection chances than on party loyalty.  Note also that he negotiated an agreement to retain his seniority, so the switch doesn’t cost him anything in that regard.

In short, this is a very significant switch for Specter’s electoral chances, not quite so significant for his votes in the Senate.  Of course, if he gets a credible challenger on the Left in Pennsylvania, then all bets are off!

If I get a chance, I’ll talk a bit about the filibuster implications in another post. But my immediate reaction is that I don’t see this having as much impact on moving the Democrats closer toward a filibuster-proof Senate as many have surmised because Specter’s vote was already driven almost entirely by electoral, and not partisan, considerations.  When it serves his purpose – as it did in the stimulus bill when he was able to leverage money to his state in exchange for his support – he will vote Democratic.  But the change in party label isn’t likely to dramatically increase his propensity to do so.