On Tuesday, several US states and localities held elections. We can learn three main lessons from the results.
Holding state and local elections in odd-numbered years is ridiculous
Holding elections in odd-numbered years makes no sense unless your goal is further increasing the power of wealthy, white, more politically knowledgeable, and older voters. These types of citizens already have many advantages in the policymaking process. It is unclear why we would want to give them a further advantage by scheduling elections when they are already much more likely to show up than other eligible voters.
To review, in the US we hold presidential elections every four years, along with House elections, a third of Senate seat elections, and gubernatorial and state legislative elections. These elections are highly publicized in the political and entertainment media, and bring with them a huge amount of political advertising and campaign mobilization. This makes many more people aware of the election and reminds them why they care about the result.
We also have "midterm" elections in even-numbered years when there is not a presidential election. In these years, the entire House of Representatives, most state legislatures, many governorships and a third of Senate seats are up for election, just like in a presidential year. Yet the lack of a presidential election reduces the amount of news coverage and campaign activity, making poorer, less politically interested, nonwhite, and younger voters less likely to show up at the polls.
All these things are exaggerated in odd-numbered years, where there are no national House and Senate elections but a very small number of state and local elections, as happened this week. Turnout among all groups of voters is very low in odd-numbered years and the demographic biases present in midterm elections are usually even more severe.
In recent years, lower turnout elections have favored Republican candidates. That was the case this year, as well. But elections prior to 2008 suggest that Democrats can still win in these lower-turnout elections if the national mood is sufficiently on their side. And in some circumstances, Democratically aligned organized groups like teachers unions have supported holding elections at lower-turnout times in order to increase their influence. But whatever the partisan implications, these lower-turnout elections reliably skew the electorate to favor more privileged groups.
All elections are becoming nationalized
This week also illustrated the increasing nationalization of local and state elections. One of the biggest elections of Tuesday was for Kentucky governor. Kentucky has voted strongly Republican in recent presidential elections, but at the same time it has frequently elected Democrats to the governorship and other statewide offices. However, this year, Tea Party–affiliated candidate Matt Bevin came from behind in the last week to defeat Democratic Attorney General Jack Conway and win the governorship. Bevin's advertising focused on tying Conway to President Obama and national Democratic Party positions.
The story was somewhat different in Virginia, which held elections for its entire state legislature. Virginia is a closely divided state that has been gradually trending toward the Democrats nationally. The nationalization (helping the Democrats) and the low turnout (helping the Republicans) appear to have canceled each other out. Despite an expensive, hard-fought campaign, in which Democratic Gov. Terry McAuliffe campaigned across the state and outside groups on both sides paid for lots of television advertising, the parties fought to a draw. The Democrats were unable to take control of the Senate (their main goal) but didn't lose seats, either. Despite several incumbents in swing seats retiring this year, no seat in the state Senate changed partisan hands. The Virginia Senate remains with a 21-19 Republican majority. The Virginia House stayed almost the same as well, with Democrats gaining only one seat.
A third example of nationalization is the State Supreme Court elections in Pennsylvania, which has been a Democratic state at the presidential level while often electing Republicans to statewide offices. On Tuesday, Democrats won all three seats on the Pennsylvania state Supreme Court that were up for election, giving them a 5-2 overall majority. This court will appoint the tie-breaking vote on the redistricting commission after the 2020 census and will soon evaluate state laws on voter ID and gun control. In Pennsylvania, the nationalization of politics appears to have trumped the effects of low turnout.
Referenda can reduce the partisan bias in legislative districts
Finally, this week's election gives clues to how state referenda could be used in the near future to change state redistricting procedures and possibly even the timing of elections. Ohio overwhelmingly passed a state referendum to reform state legislative redistricting in order to make it much harder to pass a redistricting plan without support from both parties in the legislature. Supporters of the measure say their next plan is to put an initiative on the ballot similarly reforming Ohio's congressional redistricting process.
After the 2010 census, Republicans used their control of many state governments to pass redistricting bills that favored their party. As a result, currently Republicans win a larger proportion of House seats than their proportion of House votes nationwide. But many states (outside of the South) allow for legislation or constitutional amendments to be passed by referendum. Referenda establishing nonpartisan redistricting commissions tend to be very popular (and the Supreme Court recently reaffirmed the constitutionality of these commissions). It is likely that liberals will try to put redistricting commissions on the ballot on many more states.
While I know of no organized group currently pushing this, I might suggest those worried about the unrepresentativeness of non-presidential year elections also pursue the referendum route when possible. In many states with referenda, it would be possible to move state and local elections away from midterm years (and especially the even lower-profile odd-numbered years) and onto presidential election years. This will be a harder fight to win because most voters have no preexisting views on timing of elections, in contrast to redistricting commissions, which attract instinctive public support. Changing election times will require more campaigning to explain the rationale (i.e., making it easier to vote by holding more elections all at the same time). Existing politicians are unlikely to change the election times that led them into office. But in states where it is possible, referenda are an avenue for election timing reform, in addition to redistricting reform. People interested in making electorates more fairly representative of disadvantaged groups should consider pursuing this.