13th annual Municipal Finance Conference


13th annual Municipal Finance Conference


A vast majority of counties showed increased Democratic support in 2018 House election

Voters cast their ballots on Election Day at Washington Mill Elementary School in Alexandria, Virginia, U.S., November 7, 2017. REUTERS/Aaron P. Bernstein - RC1FB1A295B0


Shifts occurred in Democratic and Republican counties, as well as suburbs and rural areas

When Democrats took 40 congressional districts from Republicans in the 2018 election, the House of Representatives experienced what many considered to be a blue wave. What does this shift mean for the 2020 presidential election? To get a better sense of this, the following analysis examines the 2018 House votes distributed across the nation’s more than 3,100 counties. This provides a more fine-grained geographic assessment of how the 2018 House support for Democrats compared with votes in the 2016 presidential election.

From this perspective, the Democratic wave is all encompassing: 83 percent of the voting population lived in counties where support for Democrats has improved since 2016. This increased Democratic support was not confined to traditional Democratic base counties. It occurred in suburbs, smaller metropolitan and rural counties, and most noticeably, in counties with concentrations of older, native-born and white residents without college degrees. Moreover, at the state level, enough states flipped from Republican majorities in the 2016 presidential election to Democratic majorities in the 2018 House elections to project a 2020 Democratic Electoral College win.

83 percent of the voting population lived in counties where support for Democrats has improved since 2016

This analysis employs recently released county-based tabulations of the 2018 House of Representatives election voting results, along with results from the 2016 presidential elections. It examines changes in “Democratic minus Republican (D-R) voting margins” between these two elections at the county level in order to determine where and by how much Democratic support has shifted over this two year period. (Note: the D-R margin is defined as the percent voting Democratic minus percent voting Republican among the all Democratic and Republican voters in the area. Positive values represent a Democratic advantage. Negative values indicate a Republican advantage.)

More than four-fifths of 2018 voters reside in counties with rising Democratic support

The nationwide D-R margin favored Democrats in both the 2016 presidential election (as Hillary Clinton won the popular vote over Donald Trump) and the combined national 2018 House of Representatives vote, with the Democratic advantage increasing between the former and latter election from 2.1 to 8.6 percent.

Democratic minus Republican voter margins

Of course, 2018 Democratic and Republican vote advantages differ across counties, as shown in Map 1.  While Republican House votes exceeded Democratic votes in more of the nation’s counties, Democratic counties tended to be larger in size, often in major urban areas. Thus, 60 percent of the nation’s voters lived in Democratic-led counties, compared with 40 percent of voters residing in counties where Republicans held the advantage.


More importantly, in a vast majority of counties—even in those won by Republicans in 2018—more voters favored Democrats in 2018 than in 2016. This can be seen in Map 2, which depicts changes in D-R margins between the 2016 presidential election and the 2018 House race. In a majority of counties (2,445 of 3,111)—irrespective of whether the final 2018 vote favored Republican or Democratic candidates—there was a positive D-R margin shift between 2016 and 2018 (meaning either a greater Democratic advantage or a smaller Republican advantage).

Map 2

At one extreme are counties in the New England states—Vermont, New Hampshire, Massachusetts, Connecticut, and Rhode Island—which voted Democratic in 2018 (shown in Map 1). Most of those counties also showed strong 2016-2018 gains in their D-R margins (shown in Map 2). At the other extreme are counties in Nebraska and Oklahoma which voted heavily Republican in 2018. As Map 2 indicates, most of the counties in those two states showed a greater D-R margin (meaning reduced Republican margin) between 2016 and 2018.

When viewed in terms of the numbers of voters residing in counties, Figure 2 indicates that 83 percent of all voters resided in counties that increased their D-R margins between 2016 and 2018—including 26 percent that increased their D-R margins by more than 10, and 57 percent that increased their margins by 0 to 9.

2018 voter populations in counties classed by change in D-R margins between 2016 (Presidential) and 2018 (House) electionsBack to top

Increased 2018 Democratic support occurred in suburbs, small metros, and rural areas.

Democrats have long done well in large urban core counties, while Republicans tend to be more popular in suburbs, small metropolitan areas, and rural communities. Using an urban typology developed by the Brookings Metropolitan Policy Program , Figure 3 shows that this characterization is valid for the 2016 presidential election, but less so for 2018 House election.

D-R voter margins, counties in different urban status categories

In both elections, urban core counties in large metropolitan areas exhibited strong positive D-R margins, while small metropolitan and outside metropolitan area counties showed negative (Republican favorable) D-R margins. Yet there was a shift between the 2016 and 2018 elections for suburban counties in large metropolitan areas from a negative to a positive D-R margin. Also, the D-R margin became more positive in large urban cores and less negative for counties outside large cores and suburbs.

As for the nation as a whole, most voters in each category resided in counties where D-R margins became more positive or less negative between the 2016 and 2018 elections (see Figure 4). This is especially notable for large suburbs, where 87 percent of voters resided in counties with increased D-R margins. For residents in both small metropolitan areas and outside metropolitan areas, that percentage was 81 percent.

Additionally, more than a quarter of suburban or small metro voters resided in counties where the D-R margin rose by more than 10. For example, in Hays County in suburban Austin, the D-R margin increased from -1 in the 2016 presidential election to +13 in the 2018 House election. Among voters residing outside metros, 37 percent resided in counties where the D-R margin rose by more than 10. While most of these heavily rural counties voted Republican in the 2018, the decline in that Republican advantage was fairly pervasive.

2018 voters residing in counties fo different urban status categoriesBack to top

Counties with “Republican” attributes showed greatest 2018 Democratic voting margin gains.

How demographically distinct are the counties that registered the greatest increases in Democratic support (or reductions in Republican support)? To assess this, it is useful to look at attributes of residents in counties that showed a sharp rise in D-R margins.

The 2016 election exit poll results made plain the attributes that differentiated Republican (Trump) voters from Democratic (Clinton) voters. While Trump voters were more commonly categorized as being whites without college degrees, older persons and native-born Americans, Clinton voters were more strongly associated as being racial minorities, persons below age 45, and foreign-born Americans.

Table 1 examines the population attributes of U.S. counties with the objective of understanding how those with the highest 2016-2018 gains in D-R margins (gains greater than 10) differ from all counties with these attributes. It makes this comparison separately for counties that voted Democratic and those that voted Republican in 2018 because, as discussed earlier, both groups exhibited increased D-R margins (or reductions in their negative D-R margins).

Demographic attributes of counties classes by voting patterns

Counties with increased D-R margins tend to have “Republican leaning” attributes, when compared with all counties: greater shares of non-college whites and persons over age 45, and smaller shares of minorities and persons who are foreign born. This occurs among both Democratic-voting and Republican-voting counties, and suggests that there was a shift toward Democratic support in counties that helped elect Donald Trump in 2016.

Back to top

2018 Democratic margins increased in states key to the 2020 election

The victorious party in the 2020 presidential election will rely on the Electoral College rather than the popular vote. A comparison of 2018 House voting results with those of the 2016 presidential election makes plain that the there is ample opportunity for a 2020 Democratic win. Map 3 depicts states where Democrats and Republicans won the cumulative state level House votes.

Map 3

It differs from the results of the 2016 presidential map wherein the Republican candidate (Trump) won more than 270 Electoral College votes, based on winning support from states such as Iowa, Wisconsin, Michigan, Pennsylvania, and Arizona. As shown in Map 3, all of those states registered Democratic advantages in their 2018 House elections. If those results hold for the 2020 election, the Democratic candidate would receive 293 electoral votes—enough to win the presidency . Moreover, in all but two states, 2018 House D-R margins showed more positive or less negative values than those for the 2016 presidential race—both in “red” Republican states and in “blue” Democratic states (download Table A). In Texas, for example, the 2016 presidential election D-R margin of -9.4 was reduced to just -3.5 in 2018.

Trump won more than 270 Electoral College votes, based on winning support from states such as Iowa, Wisconsin, Michigan, Pennsylvania, and Arizona. All of those states registered Democratic advantages in their 2018 House elections.

Back to top

What does this mean for 2020?

To be sure, these 2016 to 2018 D-R margin comparisons are suggestive at best. That is, comparisons of voting margins from the 2016 presidential elections with those for the 2018 House election—at the county and state levels—conflate support for two national candidates in the former election with that of a myriad of candidates in the latter. Still, many have argued that the 2018 House elections were a referendum on President Trump. If this is the case, then the broad shifts toward greater Democratic support—spilling over into a vast majority of Trump-won counties—could be ripe for harvesting by the right Democratic challenger to Trump in 2020.

Back to top

Note: Thanks to Cathy Sun and John P. DeWitt of the University of Michigan Population Studies Center for programming and data preparation assistance.


  • Footnotes
    1. The 2018 county level House of Representative voting data were drawn from David Leip’s Atlas of US Presidential Elections. Because 2018 county election results are not available for the Alaska and the District of Columbia, comparisons with 2016 election results do not include those counties and district.
    2. The Brookings urban classification of counties as used here represents: Urban Core (counties that are at least 95 percent urbanized within the 100 largest metropolitan areas); Large Suburbs (other counties within the 100 largest metropolitan areas); Small Metro (counties in other metropolitan areas); Other Metro (nonmetropolitan counties). These four categories represent 31 percent, 36 percent, 19 percent, and 14 percent, respectively, of the U.S. population.
    3. Estimates in this analysis of potential Electoral College votes for 2020 exclude Washington, D.C.’s 3 allocated votes, as this analysis is based on 2018 House of Representative election results, which does not include voting members for Washington, D.C.