New Vectors of Uncertainty
With at least a half-dozen toss-up states and changes to the voting status quo-through larger mail-in & absentee ballot use, reductions in mail sorting machines, and changes to voting locations, the chances of a confirmed outcome of the 2020 US Presidential election may rest in the balance of unknown for several days, possibly longer. Multiple factors affect delays in the tallying procedures; notably the reduced number of mail sorting machines and mail in boxes in August 2020 combined with a substantive increase in the number of mail-in/absentee ballots requested for this election.
By way of comparison, almost 93 million mail-in ballots have been requested for the 2020 election, in 2016 around 50 million were recorded and 15 states have broken record for absentee ballot requests, among them PA & OH. Mail-in ballots are requested by those living in their voting districts, previous to the pandemic only about 1 in 3 voters did not present at the polls. With 93 million mail-in ballot requests[1] and a current voting population around 200 million, approximately 47% of American will cast votes outside the polls. A substantial increase in comparison to 2016, requiring a significant increase in labor and coordination to be logistically managed in a short period of several weeks. As a reminder, mail-in ballots may arrive in the days following the election but must be postmarked no later than the election day. Democrats appear to be out requesting mail-in ballots, for example in PA and FL, but, again, they will be returned via mail. Requesting and mailing in a ballot, effectively adds a step relative to in-person voting.
There are more paths to a Democratic win than a Republican win based on simulations developed from the aggregation of various polls, such as the work of Nate Silver and colleagues. Presidential candidates compete to obtain the 270 electoral college votes required to win the college and be subsequently nominated as President. Twenty-two different scenarios are under consideration including 19 paths to a Democratic presidency. A renewed Republican mandate requires FL, OH, GA, NC[2], as well as TX with AK both favored to go Republican. In particular, the coveted swing states of Ohio and Florida are necessary for a Trump win[3]. Yet, internal dynamics within each state complicate predictive modelling, this is combined with systemic level concerns due to both natural and exogenous factors exerting countervailing forces. Several systemic factors are examined: changing demography of voters, turnout among young adults in the last several presidential elections, and the direct effects of COVID19 on each party’s voter base.
This analysis discusses the particularities of Florida and Ohio with an emphasis on cross-cutting factors contributing to uncertainty. Florida and Ohio are led by Republican governors facing complaints over the pandemic. Ohio also experienced a sharp increase in unemployment from 6% to 16% and has only recovered to 12%. Florida’s tourism industry has suffered major economic losses (as much as 35%) since the pandemic as well. Moreover, Florida ranks fifth nationally in income inequality. Both states serve as litmus tests for the possibilities and problems candidates face in winning 2020 election without minimal contention. Trump cannot win without both states indicating their centrality to a second term. This commentary continues with an examination of political uncertainties and bilateral regularities during a probable fragile period, followed by final points for reflection.
System Level Sources of Uncertainty
Turnout by different categories of demographic groups as they cut across multiple social cleavages complicate predictions on an aggregate scale. Three distinct factors are worth considering as confounding factors; how the changing electoral demography is shaped by race in the US, some challenges with predicting the turnout among those younger than 30, and the mitigated effect of COVID19 on the electoral bases of each party.
The last election in 2018 indicated a change in the patterns of participation transcending across racial lines. “Turnout typically falls for all voter groups in midterm elections compared with the previous presidential race, but that falloff was much smaller than usual last year. Moreover, while turnout surged across virtually all groups, it increased most sharply among the voters who historically have participated at the lowest levels. For instance, the Census Bureau reported that turnout among voters under 30 last year jumped to about 36 percent of eligible voters, compared with just 20 percent in 2014. That still left young people far behind the turnout rate among seniors, about two-thirds of whom voted, but their rate of increase from the previous election was much greater. Similarly, the Census Bureau found that the turnout rate in 2018 increased more for Latinos and Asian Americans than it did for white people.” (Brownstein, 2019). Increasing turnout among young (under 30), non-Caucasians, favors the Democratic party, however seniors remain a consistently stable block turns out to vote for conservatives more than liberals.
Additionally, there is the continual concern about youth turnout on election day, youth turnout in 2016 as expected from the two previous elections. Yet, Democrats lost support among African-American and Latino youth. Recalling the ‘youth electorate’ is more racially diverse than the ‘over-30 portion of the electorate’ at 61% compared to 73% White. Effectively, not maintaining the consistent support of the younger diverse voter is problematic as a long-term strategy for Democrats. The liberal-ticket faces challenges engaging those voters given Harris’ record as a prosecutor who enforces policies biased against minorities. Population shifts to TX by minorities after Hurricane Katrina have contributed to shifting the state closer to purple and both TX and FL are among the top 8 states in population growth since 2000. In addition, it should be noted white college educated women, increased as a proportion of the voting electorate favoring democratic candidates. If young Americans participate in the 2020 election, it should increase the chances of a Democratic win.
Population demography shifted due to the COVID19 pandemic. With hundreds of thousands of deaths, Mayer & Schintler argued in August 2020,
“overall, the death effect seems to give a modest boost to Republican chances. This is because Black and Hispanic voters favor Democrats at much higher rates than the white elderly favor Republicans. So, for every 10 deaths among African American voters, the Democrats lose eight or nine votes, while the Republicans lose only one or two. Among the white elderly, when 10 die, the Republicans lose six votes, and the Democrats lose four. Even though the white elderly are much more reliable voters, the turnout differential doesn’t come close to the gap in partisanship. Consequently, nationwide, we predict a roughly 22,000-vote a Trump vote.”
The pandemic has been comparatively more fatal for minorities than white people for example, 1 in every 920 Black Americans have died compared to 1 in every 1840 White Americans. More succinctly, the rate of mortality for Hispanic and Black American age adjusted is 3.2 times higher than for White Americans and “nationwide, Black Americans have experienced 20.8% of all deaths of known races, but represent 12.4% of the population.”
National level sources of uncertainty will have mixed effects on partisan participation. The liberals need to mobilize a solid vote from its younger supporters and those representing the diversity in the party by engaging them sufficiently in campaign issues of concern such as education, housing, and health care without provoking a cleavage on higher education versus public education or rural versus urban needs. The conservatives are positioned to benefit from strong consistent turnout among seniors, and, ironically, positive externalities from COVID19 net death are expected concerning support. Unmarried college educated females are not the stronghold of the conservatives, married (white) women and those without college educations remain distinct pools of support. For example, women now comprise 50.2% of the college-educated labor force, up from 45.1% in 2000 (Fry 2019). However, with more females of all races becoming college educated and deciding not to marry, those demographic pools are suffering from desertification.
Table 1: Confounding electoral factors and expected political party favored
Changing racial diversity of electorate Democratic
Youth voting turnout strength Democratic
COVID19 fatality effects (predicted) Republican
FL- liberalizing suffrage (ex-felons) Democratic
FL- increased Latino registration & turnout Democratic
FL- changed voter registrations (intra-US) Republican
OH- youth registration decline Republican
OH- health care & COVID19 management TIE
State Level Sources of Uncertainty
While there are national tendencies in motion, swing states provide dynamic environments for understanding the compound effects of the sources of uncertainty across differing levels of hierarchy. Universal indirect suffrage in US presidential elections constrains state-level votes to be filtered through the electoral college, where the final selection is made. State level dynamics can both facilitate or complicate voter participation and voter eligibility. Sub-federal level dynamics may play a mitigating, competing, or reinforcing role comparative to national dynamics conditioned on legislation at a state-level.
The two states under examination in this section are not only swing states but currently lead by Republican governors. Distinctly, they serve as tests of how the current administration managed the pandemic and potential economic recovery. Florida experienced deep economic, demographic losses from the pandemic but its demographic diversity is distributed throughout the state, whereas Ohio offers a moderate level of diversity, centered in the northern portion of the state[4]. As a global tourism destination, Florida has decent chances of recovering within a few years economically from this pandemic crisis, yet the future economic vitality is not so clear for Ohio. Ohio offers a state deeply affected by the pandemic with job loss and difficulty recovering. Essentially, Ohio represents past cleavages and Florida represents future possible cleavages. Both represent difficult flips from the partisan perspective, again suggesting a need to focus attention.
Cleavages on race and partisanship issues are large in Ohio representing both a rural-urban cleavage and, with the exception of Cincinnati, a north-south cleavage as the southern Ohio border sits on the extended Mason-Dixon line. Kentucky bordering the south of Ohio is solidly conservative territory and Mitch McConnell, first elected in 1984, is the longest serving senator in Kentucky history and upon election to his 7th term he will move into the top 25 longest serving Senators in US history. Ohio’s distinct geographic placement permits a unique view contrasting both cleavages due to the juncture of most racialized individuals located in the northern portion of the state. Only 25% of its congressional districts are held by Democrats, those individuals hold the urban areas, and, notably, the four individuals include the only non-males and non-Whites among the whole delegation. Ohio experienced high job losses from NAFTA generating a fertile ground for protectionist and populist discourses favoring the conservative base. Due to the complexities of the demographic terrain in Ohio, one must parse issues based on party positions in the local context to understand central tendencies. Second, Ohio youth turnout is a concern based on published research; youth failure to register is a critical signal of portending failure to participate by voting (Democratic) in the 2020 election. Finally, among states surveyed in Sept 2020 it was among those most critical of the healthcare support received for COVID but MOST supportive of Trump among all states in the survey. Ohio represents territory to lose for Republicans, i.e. the party should be able to retain it, while Florida inversely offers possible to claim for the Democratic party as it among the closest to the edge of being flipped blue.
A poll done by Your Voice Ohio/Bliss Institute asked Ohioans to rank 16 issues in order of importance. The differences along gender and racial cleavages demonstrate the importance of parsing issues relative to demography. Older individuals (white & non-white) converged on COVID19 as the top priority followed by the economy and healthcare. White more educated voters highlighted the economy and health care. Income inequality, racism and education ranked top for females. Non-White voters ranked racism, criminal justice and addiction as most important. Finally, the young are heterogeneous concerning their rankings of issues, clustering on racism, education, criminal justice, mental health, environment, social services & addiction. Ohio survey data confirm the old and young do not have any shared priorities but the young, Non-white and females cross over on multiple issues. Immigration ranks last as an issue of importance in this sample of 1037 registered Ohio voters, as married male higher income voters mark it as important. The salience of cross-cutting cleavages requires electoral participation by the young and non-White voters to produce a Democratic win. An August study confirmed youth voter registration was down by 40% in OH compared to in 2016, which favors a Republican win. Nonetheless, one key issue could exert critical countervailing pressure, health care. Ohio voters are almost perfectly divided on whether Trump or Biden are most likely to address both the public health needs and economic costs of COVID-19. For these reasons, in the absence of significant youth/diverse turnout, it is expected Ohio will remain under Republican control. In fact, Ohio is the only one of 10 battleground states where the candidates are determined as equally competent for managing the pandemic crisis, Trump has a slight edge, within the margin of error (48% vs 45%).
The issue of candidates and COVID19, more sharply divides Florida voters (59% vs 35%). The Democrats are favored. When it comes to diversity in the representation from Florida, again the Democratic party offers 11 of 13 candidates from under represented groups among 27 representatives. Florida’s diversity includes more Hispanic Americans relative to Ohio and that group has grown 9% over the last decade. Pew Research notes they comprise 17% of all registered voters, and critically, Latinos register as Democratic or Independent at a rate of nearly 3 to 1 relative to registering as Republican. Yet, “the recent trend in party registration among Hispanic Floridians stands in contrast to the pattern among all Floridians. More Florida residents have registered to vote as Republicans than Democrats since 2016” (Bustamante 2020). That notwithstanding there is a significant state-level factor unique to Florida, changes in the suffrage rights of ex-felons. Amendment 4 was passed in 2018 and has been implemented requiring former felons have met all sentencing conditions (i.e. the time incarcerated plus proof of paying any fees or financial judgements required in their case). It is difficult to estimate how many individuals will participate in 2020 but data indicate as many as 1.5 million Floridians, most non-White, now have a “clearer” path to voting. Those individuals are anticipated Democratic voters. Finally, it must be noted the last state-wide election resulted in the competitive Democratic loss of a minority candidate.
The forces at work on the political terrain in both Ohio and Florida are both dynamic and multilevel. Factors in Ohio show cleavages on gender, race, and age. In contrast, Floridians are divided more clearly upon racial and partisan lines with solid diverse turnout serving as a precondition for any swing towards liberal. The uncertainties on the political ground surrounding voter turnout strongly point to an examination of the compound effects of both system level and state level factors on predictions based on the information available about turnout expectations.
Uncertainties on the Political Ground
Predicting US Presidential election outcomes requires understanding both the aforementioned factors in combination with turnout expectations. Therefore, compound voting effects = system + state level factors, conditioned on expected turnout. Expectations for turnout among the youth and diverse groups are higher in Florida than in Ohio. If such turnout materialises (again the true effect of suffrage enlargement to ex-felons is only predicted), it should help to counteract the larger Republican registrations and the COVID19 gain the party is predicted. Participation is required for Florida to flip to the Democratic party. There are no winning scenarios for the Republicans if Florida is not defended. In contrast, Democrats can produce two winning scenarios without either state. Moreover, they benefit from 7 additional winning scenarios if they capture either state with a combination of two of the following swing states: Georgia, North Carolina & Arizona.
Final Points – Delays, Recounts, & Possible Bilateral Diplomatic Shenanigans?
This analysis does not address the challenges facing states with increased mail-in/absentee voting when it comes to tabulating results on election night. Logistical obstacles combined with competitive races in multiple swing states could produce possible delays. Indeed, litigation is underway in multiple states concerning the inequitable effects of changes made to voting due to the pandemic or other activities.
Nonetheless, looking inside at national and state level factors combined with expectations turnout, the compound effects to result in a :
FL = Democratic (flip); OH = Republican (defend)à 7 Democratic winning scenarios from that prediction all involve two of 3 swing states on the edge (GA, AZ, or NC) to win, but the fact remains Republicans must defend both FL & OH to produce a win, whereas Democrats can lose both while retaining two scenarios favors a vote convergence resulting in a liberal victory for the presidency from 2020-2024.
Thus, the Democratic party is favored to win the 2020 elections after all votes are tallied. Given the state of racial tensions and the consistent “othering” discourse of Trump throughout the elections, there is a non-zero probability of protest, scenes of intimidations to circulate about polling locations and possible violence in some state. Instances of rioting could emerge. A Democratic win may come at the cost of exacerbating fundamental cleavages during a period of crisis that themselves could undermine the democracy the party seeks to protect. If Trump rejects results outright insisting on litigation pursued at state and federal levels, it will contribute to tensions. Finally, the conditions under which Trump exits the Executive physically may play a role; signaling “stand down, but stand by” to militarised supporters is a contradictory signal at best; at worst, an outright threat to the peaceful transfer of democracy over the next several months in the US.
Recommendations for Canada & Other US Partners
Duck and cover until the election results are confirmed.
[1] Mail-in votes are regularly tabulated, however in some states absentee ballots are only recorded if the distance between candidates is sufficiently small to warrant opening them.
[2] In June, Republicans & Democrats were less than one point away in North Carolina, Trump visited NC a half-dozen times but Biden was more absent; making his second visit in September after seven months of absence. Accessed: 19 Oct 2020.
[3] GA and OH are worth similar numbers of electoral votes finding themselves on the edge of swinging red or blue. Some small probability exists Texas will swing blue with its large allocation of 38 electoral votes with 3 points dividing candidates. FL’s 29 coveted electoral votes make it a required win for Republicans though multiple winning scenarios exist for Democrats without FL.
[4] OH, one of the “true” swing states, failed to predict the Presidential outcome once since 1944, voting Nixon over Kennedy (1960).
Comments are closed.