Quinnipiac’s recent national poll made quite a splash when it reported a Democratic lead of 16 points on the generic Congressional ballot question. Indeed, conventional wisdom – as well as historical trends – suggests that Democrats, as the party opposite that of the new president and in the Congressional minority, are poised to take back one or both houses of Congress in the 2018 midterm elections. But is it too soon to jump to this conclusion? A look at the polling data from cycles past suggests it may be, and it should temper Democrats’ enthusiasm about the prospect of a wave election.
The current Real Clear Politics average of polls for the 2018 generic Congressional vote gives Democrats a fairly significant seven-point lead over Republicans. To compare apples to apples, I calculated the RCP average for each of the past six election cycles using only those polls that were conducted up to eighteen months out from Election Day (roughly up until Memorial Day of the preceding year). I also compare the RCP average of each party’s generic vote share to the actual election results (also made available by RCP). Here is the table:
As a note of caution, the 2016 average only encompasses two surveys. All others include at least ten surveys.
As the table indicates, generic vote averages eighteen months out from Election Day are simply not reliable indicators of either party’s electoral fortunes. Nearly all of these averages favored Democratic prospects, most notably in the midterm elections of 2010 and 2014, which of course turned out to be big years for Congressional Republicans. The swings between the 2010 and 2014 RCP average spreads and the spread of the actual results are stark. Only the averages for 2008 and 2006 – by contrast, big years for Democrats – were borne out in the actual election results. Indeed, in just one of the three midterm elections included here (albeit a small sample size) was polling eighteen months out a reliable indicator of the results: the 2006 Democratic sweep.
The discrepancy between what these early averages “predict” and the actual election results, particularly in the 2010 & 2014 midterms, is not necessarily due to polling “misses” that we’ve all been accustomed to hearing as of late. The 2010 GOP takeover the House was due in large part to the backlash over the passage of the Affordable Care Act that occurred over a year after these polls were taken; likewise, the GOP’s successes in 2014 were boosted by the October 2013 hiccups with the ACA’s launch – obviously missed in the May 2013 polling. Similarly, the Mark Foley scandal late in the 2006 midterms, as well as the Bush administration’s response to Hurricane Katrina in the autumn of 2005, helped to put Democrats further over the top in a cycle in which voters were already angry over the administration’s strategy in Iraq.
The Democrats also face another obstacle besides the high expectation trap engendered by these early polling averages: Republicans’ structural advantage in the House isn’t going away anytime soon. Republican Congressional candidates received 49%, 51%, and 48% of the vote nationally in the 2016, 2014, and 2012 elections, respectively, yet won 54%, 57%, and 55%, respectively, of House seats. To flip the House, Democrats need to meet or exceed their current +7 advantage on the generic ballot. This is no easy task given this Republican advantage, due in part to more GOP control over the redistricting process at the state level.
To recap: the general unpredictability of politics and the likelihood that external factors will arise between now and November 2018 mean that the Democrats’ current RCP average of +7 could very well end up being illusory. Take into account the GOP’s ability to win more seats than their national vote share would suggest, and Democratic hopes diminish further.