For a while, politicos assumed that lower turnout favored Republicans. After all, the GOP excelled in the low-turnout midterms of 2010 and 2014. The consequences of this can be found in the partisan fights over voting access, most recently regarding mail-in ballots.
2020, however, seemed to shatter this theory, with sky-high turnout resulting only in a modest Democratic victory. While Democratic turnout surged, the influx of low-propensity Trump voters beat back the wave. The theory then took another hit in the Senate runoffs in Georgia. While overall turnout was down, a relative drop in rural white turnout led to the Democrats beating Biden’s statewide margin.
Thus, a new theory of turnout emerged. The Republican coalition was now filled with low-propensity rural voters that only surge in turnout for Trump. Meanwhile, Democrats gained high-propensity suburban voters that gave them an advantage in lower-turnout elections. Some have taken this theory a step further, that this advantage could overcome the midterm backlash. Therefore, it’s worth estimating just how large a turnout advantage the new coalitions grant.
To start off, let’s explore the following hypothetical: what if the 2020 elections had midterm-like turnout? How much larger would Biden’s victory have been? To estimate this, we’ll take Biden’s margin in each county and apply them to the turnout in a previous midterm.
For example, Biden won Dane County, WI by 53%. However, instead of using the 2020 turnout of ~345k voters (netting ~180k votes), we’ll use the 2018 WI Governor turnout of ~295k voters (netting ~155k votes). This will be applied to every county in the state, with the counties serving as imperfect proxies for demographics like education, race, and density.
Applying 2020 margins to the 2018 gubernatorial race, we get a Biden win by around 1.6%. This is an increase of 1% over 2020. As seen above, this is because 2020 had disproportionately higher rural turnout than 2018.
In comparison, Obama would have only won statewide by 6.8%, a 0.1% decrease from 2012. Thus, the Democrats’ new coalition grants them about a 1% bump over their 2012 coalition. (Applying this to 2010 and 2014 WI gubernatorial elections also result in a ~1% bump).
Of course, Wisconsin is only one state – this advantage might differ greatly somewhere else. Thus, this calculation was repeated for the 2022 Senate battlegrounds of Georgia, Pennsylvania, and North Carolina. These were chosen because their counties were small enough to repeat this experiment.
(For Georgia and Pennsylvania, the gubernatorial elections were used. For North Carolina, the 2018 state Supreme Court election and the 2010 and 2014 Senate elections were used).
As seen above, the turnout advantage from changing coalitions is relatively modest. For North Carolina, the advantage is almost non-existent, and it’s only about 1% in Pennsylvania and Wisconsin. However, Georgia Democrats are somewhat well-suited to benefit, gaining about a 2% advantage over their 2012 coalition. Keep in mind this assumes the % margin stays the same for each county. That is not what happened in the runoffs.
Obviously, there are problems with this method of estimation. Some counties house very different demographic groups whose turnout patterns differ greatly. The most notable example of this is Milwaukee County, where one part is heavily black and urban and another part is whiter and suburban. In addition, certain voters within a county may turn out at higher numbers relative to others. For example, a county may be more Republican in a midterm because the county’s white voters turned out disproportionately. Finally, testing results on older election turnout will overlook population change. While I believe the effects of these are small, they are reasons to account for more uncertainty.
Therefore, let’s give Democrats the benefit of the doubt and say the advantage could be as high as 3%. Democrats still would not have been able to rely on this alone to avoid the midterm backlash. In 2010, the House popular vote would still have been at least R+4. In 2014, the House popular vote would still have been at least R+3.
The Virginia Special Election
Finally, there is another recent event that is relevant to this topic. The Georgia runoffs weren’t the only elections taking place that day. In Virginia, there was a special election to fill the vacant House of Delegates seat in HD-02. This North Virginia district is 30% black and very suburban. The previous incumbent, Jennifer Carroll Foy (D), won reelection by 22% in 2019, and Biden likely carried HD-02 by ~30%.
Despite all that, Democrats only won the election by 3% due to a slump in Democratic turnout. Yes, this may only be one low-turnout special election. However, it shows that, even in a suburban district, demographics alone do not generate enough turnout. In fact, while black turnout was relatively high in the Georgia runoffs, it cratered for the HD-02 special.
Altogether, while the Democrats’ new coalition seems to give them a turnout advantage, it alone cannot counter the expected midterm backlash. Obviously, there is no guarantee that 2022 will be a Republican-leaning year. After all, voters rewarded George W. Bush with a favorable 2002 midterm in response to 9/11. Similarly, a fantastic handling of the pandemic may help Biden out as well.
However, coalitions alone shouldn’t change one’s prior about the hundred-year-old rule of midterms. The 2022 midterm, no matter the turnout, will likely still be favorable to the opposition party.