Partisan success in House districts is significantly influenced by the composition of districts and the resulting partisan voting inclinations. Districts that are largely White and have lower population density vote Republican and those that are highly urban and dominated by non-Whites vote Democratic. That continued to play out in 2014. Although the 2014 House elections produced the highest number of Republican seats since 1928, the changes from 2012 to 2014 look less dramatic when assessed by district composition and partisan voting tendencies.
The fun of reading post-election analyses is the wide array of conclusions presented. Some argue that Democrats in the House have been banished  and could stay that way for some time.  It was a “wave” election that has ensconced Republicans in the House for some time.  Others argue that what determined the outcomes was turnout. In that view, we are just witnessing the whipsaw effect of high and then low turnout elections, so what happened in 2014 will probably have little relevance for 2016 and particularly the presidential election. 
While pundits offer many interesting and plausible interpretations of what the results reflect and mean for the future, party consultants are probably taking another approach. They recognize the national importance of short-term factors such as turnout, state of the economy and presidential ratings, but they also focus on basic questions about districts: where did the party lose, what are the prospects for regaining seats in those districts, and are there districts that offer possibilities for gains in 2016? As parties and candidates know, the composition of House districts has a significant impact on the possibility that a party can win a seat. Demography may not be destiny, but it sets probabilities that candidates must work with.
That premise has gotten less attention in recent years as individual-level analyses of survey data have established that partisan identification and ideology are playing a greater role in shaping voting decisions and perceptions of the opposing party.  Party analysts and candidates value and draw upon those findings, but their primary concern is still with the composition of districts. How does the district vote? Their best guide is often what percentage votes Democratic or Republican in presidential elections.
This issue of district composition has become more important in recent decades because of three major changes. First, the nation’s population has become more non-White. That demographic is an important indicator of the likely partisan outcome within a district. Second, the distribution of non-White population has shifted and is now more urban. The combination of more districts that are urban and heavily non-White has created districts that have a high probability of being Democratic. At the same time there are many districts that continue to be heavily White and rural/suburban and they have a high probability of electing Republicans.  Third, a lengthy realignment has created a situation in which presidential and House results increasingly coincide. The 2012 election produced the smallest percentage of split outcomes since the 1950s. The partisanship of districts now has a major impact on how both presidential and House candidates fare.
This essay seeks to pursue three concerns involving distributions and partisan outcomes. How have the three trends mentioned above played out? How much were the 2014 results shaped by these demographic factors? What are the prospects for each party heading into the 2016 elections?
As has been widely noted, the composition of the American population has been shifting in recent decades. There has been a considerable influx of immigrants, both legal and illegal. When the census is taken, individuals can designate whether they are White or non-White, or in recent years, mixed. The composition of House districts can then be determined.  Data are also available on the number of people living in a district and the square miles of a district, making it possible to determine the population density of a district. Percent urban is not useful because the threshold to be designated urban is low and does not distinguish cities such as New York City from smaller cities that are also designated as urban.
The presence of non-Whites and density are, to be sure, crude indicators of political dispositions, but they are important indicators of likely partisan voting. Non-Whites regularly vote strongly for Democrats and Whites increasingly go strongly for Republicans. High density urban areas have a long history of lower incomes and of seeing a greater desire for government programs to address the causes and consequences of inequality.
Table 1 presents the distribution of House districts for these two indicators from the 1960s through the 2012 reapportionment. Over time there has been a gradual increase in the presence of districts with relatively high density.  In the 1960s, one-half of districts were in the two relatively low density categories and about 25% were in the two relatively high density categories. By 2012, 40% were in the two low categories and about 40% were in the two high categories.
The major change has been in the presence of districts with relatively high percentages of non-Whites. In the 1960s, 65% of districts contained <10% of non-Whites. Only 13% of districts had a composition of 30% or more of non-Whites. By the 1980s, largely White districts were beginning to decline, and by 2012, only 15% of districts were 90% or more White. The percentage of districts with 30% or more of non-Whites is now 32, a considerable rise from the 13% in the 1960s. Now 55% of House districts are comprised of 20% or more of non-Whites, creating a potential Democratic base to build upon.
This increase in the presence of non-Whites is important, of course, because they vote heavily Democratic. But the increase is also politically significant because of a shift in the geographic location of non-Whites. Urban areas have been more Democratic since the 1930s.  The increase of non-Whites in American society has also been accompanied by their increasing presence in urban areas. The correlation between the percentage non-White and district density was 0.14 in the 1950s, 0.22 in the 1960s, 0.47 in the 2000s, and 0.41 following the 2012 reapportionment.
It is the joint distribution of these two traits that is important in politics. While academic research tends to want to find the independent effect of traits, in politics it is the combined effects of traits that create voting tendencies.  When a district is largely White and rural, the probability of voting Republican is high. When a district is largely non-White and urban, the probability of voting Democrat is high. Figures 1
indicate how the joint distribution of density and the presence of non-Whites have changed over time. In the 1960s, most districts were in the row of 0–9% non-White. Districts with a relatively high percentage of non-Whites were more likely to be rural than urban, reflecting the historic presence of Blacks in Southern rural districts.
By 2012, there was both a redistribution of non-Whites to urban areas and an influx of immigrants to these areas. As noted earlier, the overall distribution of districts by density has not changed dramatically from the 1960s to 2012. There are fewer low-density (rural) districts, but not major increases in high-density districts. What did change is that high-density districts now are very likely to have a relatively high presence of non-Whites. The nation has become more diverse racially and non-Whites are now relatively concentrated in urban areas.
This change has produced a significant decline in those districts that are essentially White and low-density. In the 1960s, there were 153 districts with 90% or more Whites and a density of <190 people per square mile. In 2014, there are 55 districts meeting both of these criteria. The much-discussed demographic changes and movement within our society are altering the distribution of types of House districts.
Does this altered political landscape have political significance? That is, do largely White and rural districts vote heavily Republican, and do high-density districts dominated by non-Whites vote Democratic? And, do these voting tendencies translate into consistent partisan voting for presidential and House results? That is, can we assume that how a district votes for one office will also occur for other offices? In prior decades, there was considerable focus on candidate-centered campaigns and elections, suggesting that candidates could separate their results from that for presidential candidates. What is the situation now?
The first issue is how demographics affect presidential voting. Figure 3
indicates the average percentage voting for Barack Obama in 2012 by district type. The largely White and rural districts in the lower left corner gave Obama about 40–41% of their vote in 2012. In the upper right corner (high density and percent non-White), his support was 15–27 points higher.
Variations in support for presidential candidates have become highly associated with the vote for House candidates, as shown in Figure 4
. That is not to argue that presidential candidates cause House results. Presidential candidates may sometimes dominate in defining a party image; other times congressional party positions can play a significant role; or voting for both offices may reflect the cumulative overlap of images of each wing of a party. As the House and presidential wings overlap, the underlying partisanship within districts will drive results for both offices.
However it occurs, in recent decades the correlation in voting results between the two offices has steadily increased.  Figure 4 indicates the increase in this relationship in recent decades. This increased correlation has also resulted in a significant decline in the percentage of House districts with split-outcomes, ones in which the party winning the House or presidential vote within the district differs.
With winner-take-all voting rules, if House results reflect underlying partisan dispositions, House partisan outcomes will be sensitive to district inclinations. Party planning, candidate recruitment, and campaign themes and organization make a difference, to be sure.  An articulate and attractive candidate who is well-financed and running against a weaker candidate is likely to be able to move the partisan outcome somewhat away from the base disposition, but the base partisan disposition within a district is still primary in shaping results. Figure 5
presents the percentage of House seats won by Democrats in 2012 by the same demographic groupings used in prior figures. Democrats had a difficult time winning in White, low density districts. In districts with high percentages of non-Whites and high density they won almost all the contests.
How much did the pattern for 2014 results differ from the pattern for 2012? As Figure 6
indicates, the results look much the same. Republicans gained seats in their core constituency, winning most seats in rural/suburban districts dominated by Whites. Democrats lost a few seats in districts more inclined to them in 2012. The result is that we have two political parties that consistently draw upon very different constituencies. To the extent that race and living conditions drive attitudes towards government, the result is polarization.
Comparing 2012 and 2014 Results
Party analysts trying to make sense of what happened in 2014 start with the political landscape they face. Table 2 presents some of the information they might draw on. They are looking for indicators of how consistently district partisanship is playing out and what they need to anticipate – and perhaps try to change. First, how many districts are likely to favor their party? If we use presidential voting results, the division of the nation is clear. Despite the widespread focus on how demographic trends are favoring Democrats, the 2012 Republican presidential candidate won 232 of the 435 districts, giving the party a good base to work with. As Nate Cohn has noted, Democrats may see things as shifting in their favor over time, but as of now, their strongest supporters are more spatially concentrated in urban areas, in effect “wasting” many of their votes in House contests. 
|# Districts||2012 Democratic Presidential Vote Within District|
|% Contested by:|
|% Democratic Candidates Running Ahead or Behind Democratic Presidential Vote:|
|Average D vote||27.0||39.6||41.8||49.7||58.2||73.7|
|% Dems won||2.6||3.2||12.5||65.7||95.2||100|
|# Dems won||3||2||7||23||40||126|
|% Contested by:|
|% Democratic Candidates Running Ahead or Behind Democratic Presidential Vote:|
|Average D vote||25.7||33.0||37.0||45.0||58.2||71.9|
|% Dems won||0.9||1.6||10.7||48.6||88.1||100|
|# Dems won||1||1||6||17||37||125|
The distribution of partisan presidential voting was very important in shaping the 2012 House results. Republicans contested a higher percentage of seats where their partisan base was strong and Democrats contested more seats where their partisan base was strong.  When Republican voting dominates a district, Democrats attract weaker candidates who raise less money and most Republican candidates dominate and run ahead of their party’s presidential vote. When the Democratic base is strong, their candidates face weaker Republican candidates, and Democrats tend to run ahead of the Democratic presidential vote. Strong candidates stay away from likely losses, which allows the candidate of the other party to do better.
The average House vote percentage won by Democrats varies systematically with the presidential vote. They win few seats in districts Mitt Romney won, and few of the seats they have in the House are from districts that Romney won in 2012. In 2012, the greater number of districts voting for Romney translated into a majority in the House. This is not to argue that presidential results make outcomes inevitable. Candidate positioning and campaigns are all part of mobilizing district dispositions and turning district inclinations into votes. The overall result is that in 2012 outcomes adhered closely to presidential voting.
The 2014 results, despite considerable talk about a “wave” election, suggest there was more of a marginal shift across the board in voting percentages and results. Republicans gained four seats in districts that voted Republican in the 2012 presidential election. In 2012, Republicans won 94.8% of the districts that voted for Romney. In 2014, they won 97.0% of these 232 seats. Their major gains were in districts that voted for Obama in 2012. In those districts, they gained 10 seats. As for Democrats, in 2012 the party won 189 of the 203 districts that voted for Obama, or 93.1% of those seats. In 2014, they won 179 of 203 or 87.7.
These changes could be read as a repudiation of the President, frustration with stagnant incomes, anxiety about international events, or it could be seen as the result of younger, less affluent, and non-White voters turning out at lower rates.  What shaped 2014, however, is not clear, despite the confident assertions of many commentators, and there is much to assess. These possibilities will be intensely analyzed and debated within each party before the 2016 elections. What is clear is that the drop-off in the Democratic vote in districts going for Obama in 2012 is small, indicating that there were many close races and small shifts are significant. Given that many of the 2012 voters are likely to return in 2016, the gains of Republicans are less secure than the party might want.
Despite the focus on declining competition that began in the 1970s, the role of marginal districts in shaping control of Congress is still relevant. Although many districts do not have close elections, we are in an era of competitive elections at the national level. Party analysts begin each election cycle with a concern for how many races were close in prior elections and where those districts are. They know there are many unknowns – the state of the economy, how party positions and images play out over 2 years, how their presidential candidate fares, who retires, who can be recruited, and many other conditions – but they begin with what happened in the prior elections.
How many outcomes were somewhat close and might shift if conditions changed? Table 3 presents information on this for 2012 and 2014. Following the 2012 election each party had a substantial number of Members who won with <55% of the vote. Democrats had 37 Members with <55%, with 10 of those in districts won by Romney. They had 63 Members with <60%, but 52 of those were in districts Obama won, so the situation was not dire. In 2014, they lost 5 of their incumbents who had <55 and were in Romney districts. Democrats ended up after 2014 with 69 Members with <60% of the vote. Of those 69, 62 are districts Obama won in 2012, so if they can presume “their” voters will return in 2016, the party can presume these Members will survive and they can focus on other districts.
|Number districts||2012 Democratic Presidential Vote Within District||Total|
For Democrats the problem is not so much the vulnerability of their remaining members. Their task is to recruit strong challengers so that they increase the percentage of seats they win in districts won by Obama in 2012. As Democrats look at these results, they first hope they can win 95% of “their” 203 districts, while beating some marginal Republicans. They must also recruit challengers to try to beat the 27 Republicans who won with <55% of the vote. That will be a serious challenge because most are in Romney districts.
The situation of the Republicans is perhaps the most intriguing. Their situation has clearly improved. In 2014, the party won the highest number of House seats since 1928. The party significantly reduced the number of Members with <60% of the vote, from 104 to 62. There are still 27 Members who received <55%, but only 12 of them are in districts won by Obama. Most of those receiving <60% are in districts won by Romney in 2012. Things seem very good for the party to assume the electorate supports them, if we ignore the issues of who voted in 2014 and who is likely to turn out in 2016.
Despite this, the party faces its own challenge. There are many who argue that the party must demonstrate that it can be conservative and responsible. Leaders must convince the more conservative members of the party to compromise so the legislation passed has a chance of being signed into law. But there are reasons for the conservatives not to compromise. The party has run on conservative themes and against President Obama’s policies. In 2013, they shut down the government and, despite all the warnings of dire consequences, they gained seats in 2014. Sixty percent of the Republican House Members who will comprise the 114th Congress have been elected in 2010 or after. This has been a time of sustained Republican opposition to President Obama. Those elected in recent years are likely to feel empowered to oppose Obama. They are a cohort that surely sees its attainment of office as an expression of sustained opposition to President Obama and Democratic policies.
Many also come from districts that Obama lost badly in 2012. In November, they filed a lawsuit against Obama’s postponing some regulations in the Affordable Care Act.  In the same month, the President announced a plan to shelter roughly 4 million illegal immigrants from deportation. Was he simply exasperated with Republican reluctance to pass an immigration bill, or is he trying to bolster the Democratic Party’s image among immigrants and provoke Republicans so that they overreact and damage their image among Hispanics and hurt the party’s presidential candidate in 2016? 
The party’s most conservative base would like to shut down government to stop the change in immigrant policies, but much of the electorate has gotten tired of the relentless conflict and gridlock in Washington. Politics is a process of parties endlessly trying to assess what the public wants and what a party can represent and win elections.  There is always a high degree of uncertainty as political principle and the practicality of winning the votes of the less partisan interact. The next 2 years will provide us with some great theater as the parties struggle to interpret and act on the public concerns they discern.
©2014 by De Gruyter