What the Ideology-Geography Partisan Shift Means for 2016

What the Ideology-Geography Partisan Shift Means for 2016

By David Byler - November 26, 2014

If there’s one thing we know for certain about the 2016 presidential primaries, it’s that electability will be front-and-center for the prospective nominees. Candidates will be promising that they can “expand the map,” picking up states that had previously eluded their party.  It’s not just Republicans who will make these types of arguments: Ready for Hillary’s Mitch Stewart recently named Indiana, Missouri, Arizona, Georgia and Arkansas as prime targets for Clinton’s potential presidential campaign.

Such speculation is somewhat premature at this point in the cycle, which hasn’t even officially begun. But if either of the eventual nominee wishes to expand the map into the opposing party’s strongholds -- as Stewart believes Clinton will -- he or she faces an uphill battle.

Voters are unlikely to split their tickets (vote for a candidate of one party at the top of the ticket and the opposing party farther down the ballot) because Democrats and Republicans are more organized around ideology and less around geography than they have been in nearly a century. This makes expanding the map more difficult than at any time in recent history. 

We can demonstrate this by examining four elections -- 1916, 1960, 1988 and 2012 -- that provide snapshots of how the parties moved from the stable, geographic divisions of the early 20th century to the (potentially) stable ideological divisions of this century.

 The election of 1916 shows how prewar party divisions fell primarily along Civil War lines rather than ideological ones. In 1960, John F. Kennedy’s tight contest with Richard Nixon demonstrates how Franklin Roosevelt’s New Deal and new Democratic coalition had strained the geographically based party system, driving the South towards Republicans and the North towards Democrats. Civil rights legislation, Nixon’s “Southern strategy” and later Ronald Reagan’s sizable victories would put further pressure on the old party order, and both major parties began to lose control of their historic “homes.” The election of 1988 provides a good example of this chaotic phase of the transition.

 By 2012, the transformation was mostly completed, and ideology more or less superseded geography as the primary determinant of partisan division. This change explains a large amount of current political map, and it serves as a significant roadblock to national candidates who want to flip their opponent’s strongholds.

Before diving more deeply into those four elections, it is important to describe how “partisan control” and “ticket splitting” have been measured. We built indexes for both categories by using the outcomes of elections at five different levels -- presidential, senatorial, gubernatorial, congressional and state legislative -- for each state for a given election year. The presidential component is just the Republican share of the two-party popular vote (third party candidates were excluded in all of the measures) in that year’s presidential contest. The senatorial component is the average Republican share of the two-party vote in the last two Senate elections. The gubernatorial component is the Republican share of the two-party vote in the last gubernatorial election. The House component is the average of the Republican share of the House popular vote in that state and the Republican share of the state’s congressional seats. Finally, the state legislative component is the average of the Republican share of state House and state Senate seats. (If this sounds familiar, it’s probably because it borrows from the national partisan control index Sean Trende and I published in September.)

The overall partisan control index is just the sum of these five components, translated and scaled to a -100 to 100 point score where high positive values correspond to large Republican majorities and large negative values represent large Democratic majorities.

The index for ticket splitting is slightly more complicated: The presidential component is multiplied by four and the sum of the other four components is subtracted from it. This is then translated onto a -100 to 100 point scale. A positive number indicates that the Republican presidential candidate outperformed his party in a given state and a negative number indicates the Democratic candidate outperformed his party. A score close to zero shows that both parties’ presidential candidates performed about as well as the rest of their respective parties in that state.

We can start tracing the progression from regional parties to more ideological parties using the election of 1916. That year, both parties had large internal, ideological divisions. The GOP contained sizable conservative and progressive factions, so it nominated Supreme Court Justice Charles Hughes -- a moderate who could hold the party together -- for president. Democrats re-nominated progressive President Woodrow Wilson, but a significant conservative element remained in the party. Only 12 years earlier, Democrats had nominated the conservative Alton Parker; only eight years later a badly divided party would turn to John W. Davis, a one-term congressman and solicitor general under Wilson, as its standard-bearer.

Despite the ideological diversity of both parties, party control fell roughly along Civil War lines.

Democrats dominated the South while Republicans led in the Northeast, Midwest and on the West Coast. The Mountain West, most of which had not been organized into states at the time of the Civil War, was split with some states favoring Republicans and others favoring Democrats.

Despite intraparty ideological divides, voters tended not to split the ticket.

Most states have a ticket-splitting index between 11 and -11 (not huge on a scale from -100 to 100). Additionally, some of the higher ticket-splitting scores in the South were a product of down-ticket Democratic dominance rather than up-ticket Republican success (and vice-versa for Democrats in the North). For instance, Virginia had a reasonably high index of ticket splitting (24.5) but Hughes won less than a third of the vote there. In that case, Hughes’s popularity was not driving up the state ticket-splitting index -- the complete absence of a state Republican Party outside of the Western mountains did.

Taken as a whole, these 1916 numbers paint a good picture of the stable, geographic partisan divisions of the prewar 20th century.

In 1960, cracks in this order had appeared. Although overall party dominance still fell close to Civil War boundaries (with exception of the Democratic-trending West Coast), Nixon seriously outperformed his party in the South.

During the Great Depression and World War II era, Roosevelt pulled the Democratic Party left on economics through his New Deal policies and began to build a more diverse, urban coalition. These changes stuck after the war ended, and as a result both conservative Southern Democrats and cosmopolitan Northern Republicans began to move away from their respective parties.   

This is an important point: After Roosevelt, ideology started pulling apart the geographically based coalitions. 

By 1960, the nation was in the chaotic middle phase of this realignment. The civil rights movement, Nixon’s “Southern strategy” and widening divide between the parties on economic issues were still pushing Northern Republicans and Southern Democrats further away from their parties, but the transition was incomplete.

Democratic majorities were on the verge of solidifying or collapsing in many states. Democrats maintained a lead in the South, but their level of control had declined significantly in the previous few decades. The Northeast had become majority Democratic, but Republicans still held substantial control in a few states.

This led to states with extremely mixed partisan control.

The table above shows how either party’s success at one level of elected office tracks its success at any other level. The specific measure is called “correlation.” Correlation is measured on a scale of -1 to 1. A highly positive correlation between two offices means that when one party wins in one office, it tends to win in the other. For example, the correlation between a party’s share in U.S. Senate races has a correlation of about 0.65 with its state legislative wins. That means that when Republicans won in Senate races, they typically did well in state legislative races and vice versa. A negative correlation would mean the opposite -- that if Democrats win in one office, then Republicans would tend to win in the other. A correlation of zero indicates that there is no real relationship.

The same table for 1916 and 1960 contains, for the most part, highly positive numbers. Democrats and Republicans tended to win similar shares down the ticket in all offices of a state.

The correlation matrix for 1988, on the other hand, contains a significant number of low correlations. This means that many states had a mixed partisan control and that Republican or Democratic wins at one level of office did not necessarily indicate much for either party’s fortunes at other levels. 

These low correlations demonstrate how messy the transition from geographic to ideological party organization was. The tug of war between ideology and geography was still underway, so geographic and down-ticket patterns of control were harder to discern.

By 2012, the situation had stabilized significantly.  

Republicans gained control of the mostly conservative South and West, while Democrats held control of the more liberal Northeast and West Coast. Partisan wins at the top of the ticket again coincided with wins further down the ticket. 

These reasonably strong correlations and decent (but not overwhelmingly high) levels of partisan control fit with what one might expect from ideologically based parties. In conservative and liberal states the Republicans and Democrats, respectively, hold significant control from the top to the bottom of the ballot, but neither party achieves the nearly complete domination Democrats achieved in the South in 1916 and 1960. Ideological minorities -- groups that used to be subsumed in the regional party -- side with the minority party (Republicans in the Northeast, Democrats in the South) and keep the majority party from having near 100 percent control.

Finally, note that ticket splitting was essentially extinct in 2012.


President Obama and Mitt Romney generally performed as well as other candidates from their respective state parties, despite the potential crossover appeal of a moderate Massachusetts governor and Obama’s ability to mobilize minorities in the South.

So the process has come full circle. Ticket splitting again declines and the success of a party is pretty consistent all the way down the ballot. The key difference is that now ideology divides the parties rather than geography, and that has serious consequences for 2016.

If these ideological lines hold two years from now (which they likely will), then both major party candidates will have trouble pulling reliable states from the other party’s column. If a Democrat wanted to flip Indiana or Arkansas (as Stewart suggests Clinton would), then he or she would likely have to move ideologically to the right. Similarly, a Republican who wanted to pull off a win in Oregon or New England may have to shift leftward. In this age of increasing polarization, neither of those options seems likely.

Finally, it should be noted that these ideological lines are not necessarily permanent. Historical circumstance could pry the parties into some new, unforeseen configuration, but given the actual data we have, we should probably proceed on the assumption that ticket splitting will be minimal.

David Byler is an elections analyst for RealClearPolitics. He can be reached at Follow him on Twitter @davidbylerRCP.

What 2014 Means for 2016
Michael Barone · December 12, 2014
Jeb Bush Keeps on Mulling 2016 Decision
Steve Peoples & Thomas Beaumont · December 15, 2014
Jeb Bush Forming PAC to Explore Presidential Bid
Scott Conroy · December 16, 2014

David Byler

Author Archive

Follow Real Clear Politics

Latest On Twitter