Deflation or defection in midterms?

The real question is: Were the results driven by an unusual set of suspects turning out or by voters shifting their allegiance? If it’s not already obvious, I’ll blow the surprise: 2010 was fundamentally about the failure to persuade, not turnout failure. 

Defining the appropriate point of comparison is critical. Some mislead by juxtaposing 2010 with 2008. Midterm turnouts are always different in kind from presidential-year turnouts, and while one could pray 2010 would somehow be different, expecting the 2008 electorate to be replicated in 2010 was silly. 


2006, the last midterm, not 2008, is the appropriate comparison. That hardly stacks the deck against Democrats. The 2006 turnout was sufficient to enable Democrats to wrest 30 seats from the GOP, along with control of the House.

Even a quick perusal of the exit polls (not perfect evidence with respect to turnout, but the best readily available) reveals that, despite all the talk of an unenthusiastic and deflated Democratic base, this year’s electorate looked very much like that of the prior midterm, when Democrats were bursting with enthusiasm. Indeed, 2010 voters were slightly less white (77 percent) than those in 2006 (79 percent). The proportion of African-Americans was one point higher in 2010 than in 2006, while Latinos constituted exactly the same proportion of the electorate in the last two midterms. Democrats increasingly rely on young voters, and under-30s were 12 percent in both midterms, though seniors provided a slightly larger share in 2010 than in 2006 (21 percent vs. 19). White evangelicals were 24 percent of the vote in 2006 and a nearly identical 25 percent in 2010, while residents of big cities were a one-point larger share in 2010 than in 2006 and small-town/rural voters were three points less prevalent in the most recent midterm than in the prior event. None of this suggests turnout was Democrats’ problem in 2010.

Indeed, only one segment of the Democratic coalition appears meaningfully underrepresented in 2010 compared to 2006 — union households. In 2006, 23 percent of voters lived in households with at least one union member, compared to 17 percent in 2010. The seeming decline in union participation cannot be explained by declining union membership — the census put union members at about 12 percent of the population in both years. Labor’s commitment was strong this cycle, so its declining proportion of the electorate remains a mystery, but even with union households giving Democratic House candidates 61 percent of their votes, this cannot be the whole story, especially as they gave Democrats a slightly heftier 64 percent in 2006. 

The real story of 2010 was the precipitous decline in support for Democrats across most segments. African-Americans proved an exception, giving Democrats a 79-point margin in 2006 and an 80-point advantage in 2010. Among whites, though, Democrats collapsed from -4 to -23, while the margin among Latinos declined from +39 points to +22. Democrats lost white women by a single point in 2006, but by 19 in 2010.

Rural voters proved a real battleground in 2006, when Democrats fell only three points behind with this segment, but lost them by 31 points this cycle.

Perhaps the most dramatic disintegration of Democratic support was among independents, who gave Democrats a resounding 18-point victory in 2006, but four years later backed GOPers by a similar 19-point margin.

Democrats lost in 2010 not because our voters failed to cast ballots, but because they deserted us once inside the voting booth.

Mellman is president of The Mellman Group and has worked for Democratic candidates and causes since 1982. Current clients include the majority leaders of both the House and Senate.