The Gender Gap Giveth, the Gender Gap Taketh Away
One of the more prominent bits of conventional wisdom this cycle is that Hillary Clinton’s advantage among women will open a huge gender gap, and that this will help her win the presidency. I don’t doubt that there will be a larger gender gap, perhaps even a historically large one. I’m less convinced that it will work to Clinton’s benefit.
There are two reasons for this. First, my sense is that many analysts overrate the pull identity politics has among women. This is more of an intuition than anything I can draw from data – we don’t really have a good dataset on women presidential candidates!
So I turn to the second point: The gender gap is an inherent, long-standing force in American politics. It does not, however, always work to Democrats’ advantage.
To see what I mean, consider the case of Mark Udall in Colorado. Udall was elected to the Senate in 2008 with 55.4 percent of the two-party vote (that is, excluding third parties from vote totals). There was a gender gap that year, but it was relatively moderate: Udall won 58 percent of the female vote and 52 percent of the male vote (two-party).
When he ran unsuccessfully for re-election in 2014, the environment was very different. Fortunately for Democrats, they saw it coming, and had time to plan accordingly. Udall decided to take a page from previous Democratic playbooks – Colorado U.S. Senate candidate Michael Bennet in 2010 and Barack Obama in 2012 – and play heavily to groups that tended to lean Democratic, emphasizing women, and single women in particular.
Udall’s strategy became something of a running joke in Colorado, with the nickname “Mark Uterus” sticking at some point. But here’s the thing: To a degree, it worked. Udall won women, with 54 percent of the two-party vote. Despite the massive shift in the national environment, he only dropped four points among women.
The problem is, the focus on women’s issues probably backfired among men. Udall plummeted from 52 percent of the male vote to just 41 percent. In the end, his overall vote share dropped seven points, similar to the shift in the national House vote of nine points. Perhaps that two points was due to his heroic efforts among female voters, but one could just as easily attribute the shift to the fact that, unlike 2008, this race stayed close until the end (unlike 2008, when Republicans wrote off the race at the end), or to the fact that Colorado introduced all-mail voting for 2014. Regardless, this illustrates that opening a massive gender gap does not necessarily work in the Democrats’ favor.
We can see this at the national level as well. The following chart plots the gender gap against Democrats’ national vote shares over the years. A negative number suggests a larger gender gap here, for visual reasons.
In every one of these elections, there has been a gender gap of varying sizes, and Democrats won the female vote in almost every election (2010 is the sole exception, when Republican narrowly won among women). Yet, if anything, a larger gender gap correlates with a worse Democratic performance in House elections (p<.001). Two of the three largest gender gaps on record – 11 points in 1994 and 10 points in 2012 and 2014 – occurred in two of the worst years for Democrats nationally. This happened because Democrats performed horribly among men, winning just 42 percent each time. In fact, if we look at the entire time series, the men’s vote is far more variable (standard deviation = 4.2) than the women’s vote (standard deviation = 2.4).
We can see this at the national level as well.
Democrats won the female vote in almost every election in the past 20 years (2010 is the sole exception, when Republican narrowly won among women). Yet, if anything, a larger gender gap correlates with a worse Democratic performance in House elections. Two of the three largest gender gaps on record – 11 points in 1994 and 10 points in 2012 and 2014 – occurred in two of the worst years for Democrats nationally. This happened because Democrats performed horribly among men, winning just 42 percent each time.
Finally, current polling is consistent with the notion that a gender gap could hurt Clinton. In the recent ABC/WashingtonPost poll, Clinton does have a significant lead among women: 16 points. But she trails because Trump trounces her among men, by 22 points. Likewise, the Fox News poll shows Clinton defeating Trump by 14 points among women, but losing men by 22 points. The NBC News poll, in which Clinton leads, actually shows a significantly smaller gap, with Clinton performing worse among women: She leads by 13 points among women while losing men by nine. The RealClearPolitics' average is here.
It is far too early to predict anything with certainty this cycle. But that includes the size of a gender gap, and who would be the beneficiary. All we can really say right now is that the gender gap giveth to Democrats, but it also taketh away.