If the Senate allowed a Merrick Garland vote, he might pass

A basic data analysis suggests that 60 senators might support Merrick Garland for the Supreme Court. Perhaps more interesting, though, is how much has changed since he was confirmed to the D.C. Circuit. 

On Wednesday, President Obama appointed Merrick Garland to replace Antonin Scalia on the Supreme Court. Senate majority leader Mitch McConnell (R) of Kentucky and Senate Judiciary Committee Chairman Chuck Grassley (R) of Iowa have said that Judge Garland will not receive a vote in the current Congress.

Despite the statements of the two senators, one of the reasons Mr. Obama selected Garland is the fact that he is relatively moderate, as indicated by this Monkey Cage post, and was thus approved by the Senate in 1997 by a vote of 76 to 23. Notably, Garland was appointed to the D.C. Circuit in that year by a Republican controlled Senate, receiving the support of a majority of Senate Republicans (32 to 23).

On the surface, this would seem to bode well for Garland’s chances of making it to the Supreme Court despite obstructionist claims.  However, much has changed since 1997.

In this context (and many like it) the biggest change since 1997 has been the increase in ideological polarization in Congress. Countless studies have documented that Democrats and Republicans have drifted further apart over the past three decades. My work on this topic argues that a major reason for the Senate’s polarization has been the increase in the number of senators who first served in the (more polarized) House of Representatives (see here and here).

A straightforward analysis can help us adjudicate between these competing expectations about Garland’s likelihood of being confirmed (and crudely estimate how many votes he might receive). Although this analysis has a number of limitations, it can give use a rough sense of what to expect.

I estimated a simple logit model of the 1997 vote to confirm Garland to the D.C. Circuit. All data came from Keith Poole’s Vote View website. The model includes just two variables: a senator’s ideology and whether that senator is from the South. Although this is a very simple model for sure, it performs well, correctly predicting 88 percent of the "yeas" and "nays" in 1997.

Based on that model’s estimates, I then predicted what would happen in the current Senate. Data on the ideology of senators in the 114th Congress also came from Poole’s Vote View website.

Figures 1 and 2 below present a senator’s predicted probability of voting for Garland (on the y-axis) by their ideology (on the x-axis).  Liberals are on the left, conservatives on the right. Figure 1 is for the actual for Garland in 1997 (105th Congress) and Figure 2 is for the simulated vote in the current Congress (114th Congress). Democrats are in blue and Republicans are in Red.

FIGURE 1

FIGURE 2

In the current Congress, the model estimates that 60 senators would vote for Garland and 40 would vote against his confirmation.  Looking at the figure, 60 senators are above the 50 percent threshold (more likely to vote for Garland than against) and 40 senators are below the 50 percent threshold.

All Democrats are predicted for Garland as are 14 Republicans. The predicted Republican yes votes are: Sens. Lamar Alexander of Tennessee, Cory Gardner of Colorado, Shelley Moore Capito of West Virginia, Jeff Flake of Arizona, Mike Rounds of South Dakota, John McCain of Arizona, John Hoeven North Dakota, Dean Heller of Nevada, Rob Portman of Ohio, Orrin Hatch of Utah, Kelly Ayotte of New Hampshire, Mark Kirk of Illinois, Lisa Murkowski of Alaska, and Susan Collins of Maine.

While this prediction may seem surprising given what we know about party polarization, the two figures reveal how the Republican Party’s shift to the right changed the landscape of Senate confirmations. Indeed, if just 14 Republicans vote for Garland that would represent a net decrease of 18 Republican votes from 20 years ago.

In the first figure above we can see how in the 105th Congress, a number of moderate Republicans have high probabilities of voting for Garland according to the model (and indeed, moderate Republicans did vote for Garland in 1997). But as new, more extreme, senators replaced these moderates, the individual probabilities of a “yes” vote drop well below 50 percent line.

Let me emphasize, again, that this is a crude analysis. It assumes that ideology and region alone shape how a senator votes on confirmations. Of course, there are electoral and strategic reasons why a senator would vote for or against a president’s Supreme Court nominee. In addition, an excellent Monkey Cage post by Kastellac, Lax, and Phillips discusses how public opinion in a senator’s state is a key piece of the puzzle.

At the minimum, however, the analysis gives use a sense of what a vote on Garland’s nomination might look like and how the landscape of Senate confirmations has been shaped by ideological polarization. Of course, there’s a good chance we will never know if this prediction is right or wrong.

Jordan Ragusa publishes his Rule 22 blog at http://rule22.wordpress.com.