At Electoral Crossroads: An Analysis of Federal Elections Since 1984 With November 2020 In Mind


“As the select assemblies for choosing the President, as well as the State legislatures who appoint the senators, will in general be composed of the most enlightened and respectable citizens, there is reason to presume that their attention and their votes will be directed to those men only who have become the most distinguished by their abilities and virtue, and in whom these people have just grounds for confidence…The inference which naturally results from these considerations is this, that the President and senators so chosen will always be of a number of those who best understand our national interests, whether considered in relation to the several States or to foreign nations, who are best able to promote these interests, and whose reputation for integrity inspires and merits confidence”.[1]

John Jay, Federalist Paper LXIV

“How best to meet [these] democratic requirements in a political unit as large as a country is, of course, enormously difficult, indeed to some extent unachievable…Clearly the requirements could not be met if the top officials of the government could set the agenda and adopt policies independently of the wishes of citizens. The only feasible solution, though it is highly imperfect, is for citizens to elect their top officials and hold them more or less accountable through elections by dismissing them, so to speak, in subsequent elections”.[2]

Robert Dahl, On Democracy

2020 – A Unique Election:

An Introduction

On Tuesday the 3rd of November 2020, the United States of America (US) will hold its federal elections. These elections will determine who holds the offices of: (a) the President, (b) 35 seats of the Senate, and, (c) all 435 seats of the House of Representatives. It has been impossible to escape discussion about the presidential election for some time, due to it being fought between veteran Democrat and previous Vice-President – Joe Biden, with his running mate Kamala Harris – and the incumbent Republican President – Donald Trump, with his current Vice-President Mike Pence. In 2016, Donald Trump won the presidential election, surprising most pundits and political scientists alike, by gaining 306 electoral college votes, 74 more than the ex-First Lady and former Secretary of State – Hillary Clinton. Since then, it has  often been claimed that the politics of the US has become discordant, where large social and political cleavages openly divide citizens between those ‘liberals’ who frequently flood the streets to protest the Trump administration on the grounds of his character or policies, and those ‘conservatives’ who counter-protest, wishing to reinforce those ‘traditional’ values that they believe the Democratic Party as embodying the erosion of.[3]

Concerning Elections

With this in mind, it is vital to remember that elections are an essential quality of democratic political systems. Without free, fair and frequent elections one would struggle to classify any given political system as democratic to even a minimal degree. Whenever we engage with differing conceptualisations of democracy, of which there are a number, the explicit reference to ‘elections’ is not always a recurring posit.[4] What does reoccur, however, is the notion that ‘citizens’ (as ‘the people’ – δήμος) engage in a process whereby their collective course of action is decided upon through some mode of deliberative arbitration of distinct judgements.

This is the first purpose of elections, and it sits within a normative realm: to settle civic disputes as to what collective action, and its ends, should entail. In representative democracies, whereby the concerns of citizens are mediated by elected officials who engage in the activity of governance, elections provide an instrumental function. Simply put, elections afford a mechanism by which individuals who seek to represent and govern are rendered the democratically legitimate basis to do so.

In a retroactive manner, affording officials their democratic legitimacy (the democratic right to rule, govern and wield power) is an exercise in public committal to civic responsibility. Expressed another way, ‘we’ cast a judgement on the collective action that ‘we’ wish our political unit to engage in; this unit being the civic body that we are a citizen-member of. With this in mind therefore, ‘we’ as a single unit of citizens, connected  by virtue of sharing some civic space, are responsible for the political action taken by our representatives precisely because ‘we’ afford them the legitimacy of holding office – at least on the condition that such action was clearly willed and declared by the individual candidate prior to their taking office.[5] Taking all of this into account, elections provide three functions in representative democracies: (1) of instrumentality, (2) of legitimation, and (3) of civic responsibility.

Concerning This Election

What makes this set of US elections so unique is not simply that ‘America Decides’ and a public judgment is made concerning the future leadership and staffing of federal representative institutions. Rather, with this set of federal elections there is so much at stake with its outcome.

During the previous election Trump was able to cultivate the long-fermenting image of the outsider who would adapt the ossifying status-quo to benefit those sections of society that had been labelled as ‘forgotten’ or ‘neglected’; polling well with the white rural working class in the mid-west and the south.[6] Indeed, this image was given credence to with Clinton’s perhaps most infamous and readily forgotten remarks of her campaign, both said within moments of one another. Here, Clinton claimed that some Trump voters could fit into a demographic that she termed the “Basket of Deplorables”, who are “racist, sexist, homophobic and xenophobic”, whilst yet consisting of “people who feel that the government has let them down, the economy has let them down, nobody cares about them, nobody cares about what happens to their lives and their futures, and they’re just desperate for change”.[7] Although many have often come to forget the latter statement, Clinton’s remarks oddly vindicated the very point she was attempting to make, i.e. that there are swathes of the US electorate that are politically disenchanted for iterations and utterances exactly like these – leading them to will a ‘draining of the swamp’, in Trumpian terms.

If we can state that the outcome of the 2016 election affirmed Clinton’s infamously forgotten utterance that a large portion of the electorate were ‘desperate for change’, the concern of this election will be the public judgement of the character underpinning such a change. Some may contend that this election will be a ‘referendum’ on the Trump administration, but what does this utterance really mean? In this, the term ‘Trump’ is a linguistic signifier for the political phenomenon that is the sum of the norms and policies that President Trump has made common-place over the course of the past four years.

Trump’s presidential term has been saturated with controversial policy. From Executive Order 13780 and the subsequent Presidential Proclamations 9645 and 9983 limiting or suspending immigration and asylum seeking from various states, to the manner in which his administration has handled the COVID-19 pandemic (with 7.3 million total cases – himself included), to his stance on the killing of George Floyd and the ensuing Black Lives Matter protests, Trump’s attempt at adapting the status-quo of US politics has brought normalised controversy to the heart of the political system.[8] This election is unique, not only because of the widespread division in the US that we are able to observe with daily events, but because it will afford the space for the electorate to judge the extent to which Trump’s controversial policies are qualitatively desired, and by extension of this, making them civically responsible for legitimating or rejecting Trumpian politics as the normative basis of the American political landscape.

Divided Government and The Two-Party System

Additionally, this election cycle is far more than just a presidential race. Come 2021 the partisan make-up of Congress may have also shifted, as all 435 seats of the lower house and 35 seats of the upper house are up for election also. As designed by the classically liberal founding fathers, power on the federal level is disseminated between three institutional bodies (the Executive, Legislative, and Judicial), each of which holds the capability to check and balance the others and thereby “power should be a check to power”, keeping tyranny at bay.[9] The US government operates along terms of constitutional self-limitation in order to uphold this notion that individual political liberty results from limiting the capability of the state to interfere in the lives of citizens through a separation of powers, following the thought of Montesquieu in his ‘The Spirit of the Laws’.[10]

 If presidential elections concern the executive branch of government, then the elections for the House of Representatives and the Senate concern the legislative branch – for control of Congress. In this mix, the United States operates with an elective bicameral legislative system, in which both the upper and lower houses are subject to full elections with differing electoral cycles in order to actualise a further self-limitation of government.

When the executive and legislative branches of government are not controlled by the same party, elective government is divided. This scenario is not uncommon, with the Nixon, Reagan, Bush (Sr.), Clinton, Bush (Jr.), Obama and Trump administrations in office whilst experiencing some varying degree of division, either with Congress as a unit or with just one of the congressional chambers. In this scenario, self-limitation can quickly become stagnation, inhibiting the passing of swift, coherent and purposeful action as a result of intra-governmental partisanship.[11]

Another aspect of the US Political system that adds to the stagnation that can be experienced by divided government concerns its party system. Every modern democracy is correspondingly a party democracy and consequently the nature and number of parties competing for governance adapts the character of the state’s political machinations, structures, processes and so on.[12] The United States operates with a Bipartisan order, often referred to as a ‘two-party system’. Typically thought of as one of the greatest political scientists to explore the relationships between parties, and how they form an evolving party system with their continual interaction, was the Italian thinker Giovanni Sartori. In his ‘Parties and Party Systems’, Sartori goes to great lengths in order to strain-out a phenomenologically accurate conceptualisation of the modern two-party system. Here, he claims that a system would function according to the following rules if it were a two-party system:

“(i) Two parties are in a position to compete for the absolute majority of seats; (ii) one of the two parties actually succeeds in winning a sufficient parliamentary majority; (iii) this party is willing to govern alone; (iv) alternation or rotation in power remains a credible expectation.”[13]

Although Sartori centres his focus on the legislative branch of government, the same four symptoms of a bipartisan system can be observed in any of the democratically electable positions or chambers of the federal government. What unifies all four is that they emphasise clear division, competition, non-collaboration, and even borderline belligerence. This comes together in determining that a two-party system is one in which only two major parties have a realistic chance of holding power.[14]

The US is often illustrated as the archetypal two-party system, in which conflict between the Republican and Democrat parties are fought at every political level imaginable – from the micro-local to the hyper-federal. As if the bipartisan system was not concrete enough, further entrenchment of this division has been formally engrained in legal regulations, encouraging a sympathetic oversight of smaller parties like the Greens or the Libertarians who struggle to even get a candidate on the ballot due to judicial requisites, and a lack of public support and funding.[15] Essentially, this means that in a sovereign federal state that champions the freedom of thought, representative democracy, the free market, meritocracy and financial competition – political capital and capability is limited to the status of a plutocratic oligopoly.

Of course, this makes it exceptionally difficult for political groups outside of this edifice to gain any political influence whatsoever, but it also adds to the potential stagnation of legislative proceedings. Amidst the ideological break with the past we have experienced under the Trump administration, pushing the Republican party further to the populist right, the divisions between these partisan monoliths at the centre of power have only entrenched themselves to a greater depth. For example, whether the reader agrees with it as a policy or not, take the construction of Trump’s proposed wall along the Mexican border. Throughout the 2016 election campaign, Trump consistently claimed that he would construct ‘a great wall’ between the US and Mexico in order to temper the number of migrants crossing the border illegally. After his inauguration, especially since 2018, this became an almost impossible task with Congress holding ‘the power of the purse’, requiring the ideologically opposed Democrat controlled lower house to pass a budget that would have permitted the wall to be built with Congressional approval. This did not happen and subsequently led to the 2018-2019 government shutdown that lasted 35 days, the longest in US history – stifling federal politics for this period. The same was the case for Barack Obama in 2013, where the then Republican controlled house blocked the budget over the Affordable Care Act (‘Obamacare’) on a logic of principle.

This is significant as it demonstrates how the two-party system adds pressure and increases the possibility of legislative stagnation in the event of divided government. With the two-party system operating in tandem with a divided government a Manichean perspective of politics blossoms and quickly festers; a loss for ‘them’ is a disguised gain for ‘us’, even if it means all federal legislative politics is put on a hiatus. This is where the two-party system and divided government begin to co-constitute one another and thus instil a sense of institutional impotency about the political system as a whole.

Of course, a number of Scholars contend that a divided government is not a limit on the efficacy of government. In his ‘Divided We Govern’, the renowned political scientist David Mayhew presented this thesis. Mayhew found that there was little difference concerning the environment of policy making between divided and unified governments, i.e. the volume of significant laws passed in periods of divided government did not considerably differ from periods of unified government.[16] This being said, perhaps Mayhew’s argument errs in locating nuance. Mayhew chooses a simple measure in order to conduct his research – the supply of legislation – and hence neglects to include into his design a measure that concentrates on the demand for such legislation.[17] The question remains therefore, does divided government stagnate legislative action that is normatively willed by the electorate, or does it simply make executive dominance of the political edifice more complex to achieve?

As it stands, the Republican party (Grand Old Party – GOP) hold the executive branch with Donald Trump as the incumbent. Equally, the GOP holds the upper house of Congress, the Senate, with a three-seat majority. The Democrats, since 2018, have held a majority in the lower house of Congress, the House of Representatives, with 35 more seats than the GOP. In this manner, the GOP has held the capability throughout the Trump administration to block and filibuster Democrat legislation arising from the lower house, and vote in favour of Trumpian policy – the best example of which was the rejection of the two articles of impeachment against Trump, drawn up and agreed upon by the lower house, where only one senator (ex-presidential candidate Mitt Romney) did not vote along party lines. Henceforth, as things stand, the US political system subsists in a condition of quasi-division where the upper-hand, institutionally speaking, rests with the GOP.

What does all of this mean? Essentially, what I am suggesting is that although a lot of emphasis has been placed on the presidential race between Trump and Biden – which is understandable as it is an election for the head of state – the outcome of the federal congressional elections will determine not only the normative content of legislative proposals in the next two years, at least, (be them increasingly Trumpian or not) but, perhaps more importantly, partisan control over the federal legislative process. This is what makes this election unique. Not only is America to decide upon who it wishes to be Head of State and which normative framework of governance to wholly legitimate, but it is also to decide which party is to control the institutional and legislative agenda of government – making its institutions capable of wholly resisting or aiding the head of state come January. This is no ordinary election; its outcomes will have both widespread tangible and normative consequences for the US. In this manner the elections on November 3rd are unique.

This Investigation

The purpose of this exploration is not to calculate, postulate nor predict which party or candidates will win these elections. This would require a methodological approach that pollsters such as YouGov or Ipsos MORI are far better at designing and constructing than any lone researcher, not least as they hold the capability to collect data that is befitting to large-n quantative studies. The concern of this investigation will be to inform the reader where the election will be fought by empirically analysing data from past elections. As each of the elections being fought on November 3rd could hold a number of potential consequences, this paper will seek to inform the reader as to (a) which states/seats may swing and as such will receive attention from the major political parties in the coming weeks, and (b) simply which seats/states are generally of interest, confusing many as to how they will swing, especially given the unique qualities of this election.

Overall, this paper will find that this election is indeed the Republican’s to lose, with vital states and seats hanging by a thin GOP majority as it stands. If the GOP was to lose just a handful of seats in the Senate, and in some states as little as 2% of the popular vote, the entire US political system could swing to the Democrats. As the GOP currently hold two of the three federal institutions elected by popular vote it is for the GOP to lose.

In order to discuss this thesis, this paper will be divided into two major chapters. The first will seek to discern where the presidential race will be fought, and to locate which of the fifty states are imperative for each of the major parties to win. Here, I will begin with a short description of the presidential electoral process before discussing trends in electoral data from US presidential elections since 1984. I have chosen to engage with such a task in order to overtly lay out the statistical context for this election. From here, the 2016 election will be analysed, and so the potential swing states will be located, placing a magnifying glass over each of these states in order to show which counties are the most likely swing. The six states this paper will identify as crucial for both parties to focus on are: Wisconsin, Pennsylvania, Minnesota, Michigan, North Carolina, and, of course, Florida. 

Once this has been achieved, my analysis will move to Congress. I will begin by focussing on the lower house of Congress – The House of Representatives. This section will, broadly speaking, be broken into four parts. The initial segment will explain the electoral process for The House of Representatives and how this differentiates with the other elections. Following this, the quantative electoral data of The House of Representatives from 1984 onwards (upholding analytic consistency with discussion of the Presidential elections) until 2018. The current state and structure of the House of Representatives will then be discussed utilising the results of the 2018 election, highlighting how this midterm election shifted the balance of power in government away from the Trump administration. Lastly, as a lone researcher attempting a project of this size cannot review the status and likelihood of outcome for all 435 seats of The House of Representatives before the election, I shall highlight a handful of seats and themes that will be of interest.

The penultimate half of this chapter, with its focus on Congress, will seek to analyse the Upper House of Congress – The Senate – in much the same manner as the Lower house. The chief difference between the two is that as only 35 seats are being decided upon, a closer reading and interpretation of what will be interesting races will be undertaken. This section will begin by explaining the electoral process for The Senate, and much like the other sections, lay out why this process is different. Then a statistical analysis attempting to discern any trends from elections prior to 2018 will be completed. Lastly, this section will discuss in greater detail the seats that are up to be voted on in November, the 2014 results, and those seats that will be close or fascinating to observe the outcomes of.

Finally, this investigation will come to a close with some concluding remarks. This conclusion will seek to summarise and retrace each step of the investigation as whole before leaving the reader with a few final thoughts.

Chapter I.

The Presidency

Trump and ‘Populism’

In 2016, Trump won by appealing to his base – the rural white working classes – and by tapping into a discontentedness with the status-quo of the US political system that had been simmering away for some time. Indeed, it are these right-wing principles that many voters and pundits alike deem to be distasteful, somewhat vulgar, or regressive even. We know that Trump is a right-wing ‘populist’ and that this won him the election, but what does this mean? Although there is much debate concerning a definition of populism, adapting and evolving with the times as all political phenomena do, ‘Populism’ can be broadly conceptualised as:

“An ideology that considers society to be ultimately separated into two homogenous and antagonistic groups, ‘the pure people’ versus ‘the corrupt elite’, and which argues that politics should be an expression of the volonté générale (general will) of the people”.[18]

In his 2019 work ‘The Far Right Today’ Cas Mudde, the scholar whose conceptualisation of populism above is often thought to be the most holistic, saw to explicate something further of this phenomenon that we can comparatively detect  across the globe in states like Brazil, with Bolsonaro, India, with Modi, Hungary, with Orban, and so the list could go on.  Interestingly, in this study, Mudde claims that although Trump holds aspects of the ‘Radical Right’ – in seeking to democratically reform the US away from its liberal foundations, but without a general scepticism of democracy – the normative shift that Trump has attempted to undertake has been endeavoured by replacing personnel and the staffing of government as opposed to overhauling the systemic institutional structure of the US; alongside this,  putting in place policies that may indeed constitute the backsliding of liberal democracy, such as actively encouraging a mistrust of the free press and media outlets that he labels as the ‘fake news’ of the elite.[19]

 Although a number of commentators often like to place Trump into the category of the ‘Alt-Right’ [Alternative Right] alongside extreme right-wing militias and ethno-nationalist groups in the US, this would be to conflate the Trump phenomenon with another that has also been quietly bubbling below the surface in the US: White Nationalism. In his timely and uniquely explanatory work into the ‘Alt-Right’, the scholar George Hawley perfectly distinguishes the Trump administration, and even Trump himself, from the ‘Alt-Right’. Here, Hawley claims that:

“We cannot accurately describe President Donald Trump as part of the Alt-Right. He is a right-wing populist, whose rhetoric has appealed to xenophobic elements of the electorate, but there is no compelling evidence that he is a white nationalist who seeks a pure white ethnostate. Saying that Trump is not part of the Alt-Right does not require endorsing his rhetoric or his administration’s policies; it is simply a recognition of the Alt-Right’s radicalism”.[20]

What Hawley reveals is that when analysing the norms shift of the Trump phenomenon, we must be careful not to conflate or confuse it with the extremities of White ethno-nationalism, even if these extreme groups express support for the Trump phenomenon.

This being said, Trump’s populistic praxis remains in concordance with Mudde’s grasp of populism generally, i.e. as an anti-elite illiberal force. Part of this praxis concerns the manner in which Trump became president with his use of data harvesting and analysis to ‘hack’ the election. In 2016, the Trump campaign formed ‘Project Alamo’ – a database of voter information that became the beating heart of the campaign’s fundraising and advertising operation. Essentially, the 2016 Trump campaign utilised social media data from Facebook, with help from the now infamous firm ‘Cambridge Analytica’, in order to generate a pooled reservoir of ‘psychographics’. Psychographics can detail any given member of the electorate’s psychological profile, how they would be likely to vote, and hence the best issues and environment in which to advertise a given candidate in order to secure the citizen’s vote.[21] 

Some may suggest that such action is not a subversion of democratic norms and principles, but has been a part of them for over a century. Although we have just begun to enter the digital age of social media and data harvesting, ‘political data’, as collected knowledge of the electorate, has played a part in the US political system for some time. In 1891, the chairman of the Republican National Committee, James Clarkson, discusses his own use of voters’ information for political gain. Clarkson claimed that:

“I had with two years of hard work, secured a list of the names of all the voters in all the important States of the North…and lists which gave the age, occupation, nativity, residence and all the other facts in each voters’ life, and had them arranged alphabetically, so that literature could be sent constantly to every voter directly, dealing with every public question and issue from the standpoint of his personal interest”.[22]

What difference is there between earlier forms of political data harvesting in the US, like the lists drawn up by Clarkson in 1891 detailing the psychological profiles of important voters and organising an array of campaign literature that would best sway that individual to vote Republican, and Trump’s Project-Alamo of 2016? Very little. Both sought to collect information that would allow their campaigns to become increasingly effective amongst key demographics and localities. There are nonetheless two rather significant variances that make all the difference: (a) scale, and (b) misinformation.

Project-Alamo harvested data on a scale befitting only the mass-politics of the times. In his discussion concerning the intersection between continually undermined democratic processes and contemporary technology, Jamie Bartlett’s ‘The People vs. Tech’ illustrates the sheer scale of Trump’s data-driven campaign. Aiming to illustrate to the reader why our democratic processes require regulations and firewalls against the politics of Project-Alamo, Bartlett shows that Trump’s campaign honed its focus on just 13.5 million key voters in 16 battleground states, ascertaining millions of data points on each voter in order to generate a ‘computerised dashboard’ which offered combinations of: TV advertising, rally locations, doors to knock on, and where to direct emails that would, overall, yield the best results.[23]

The scale of such an operation, one that sought (and succeeded) to morph an entire candidate’s internal bureaucratic framework into a data driven machine, begs the question about the extent to which such action undermines the free will of the voter. If the millions of data points collected concerning a single individual enables a candidate’s bureaucratic edifice to generate their psychological profile, and then calculate the best manner in which to campaign to their locality or demographic, what the individual voter thinks, wills and believes is irrelevant; what becomes relevant is the manner in which that voter is determined to vote for the candidate in question with their own psychology and political preferences used to disturb their traditional beliefs, principles and previous voting patterns. Project-Alamo was not an exercise in undermining democracy as a whole, but it did seek to curtail the liberal notion that the electorate should be freely persuaded to choose a specific candidate – not determined with their own psychology against them.

Alongside this, the Trump campaign utilised its data harvesting to propagate the most effective ‘misinformation’, i.e. that which would yield the greatest pay-off. In his television advertising, the ‘liberal elite’ were continually claimed to be conspiring against the US people in a number of ways, with Hillary Clinton at the apex of this treachery. In one somewhat relatively moderate advert, for example, the Trump campaign asserted that:

 “In Hillary Clinton’s America the system stays rigged against Americans, Syrian refugees flood in [appearing on screen is a ‘quote’ from Clinton stating that ‘U.S. should take 65,000  Syrian refugees’], illegal immigrants convicted of committing crimes get to stay – collecting social security benefits, skipping the line – our borders are open, its more of the same but worse. Donald Trump’s America is secure, terrorists and dangerous criminals kept out, the border secure, our families safe. Change that makes America safe again.”[24]

As this ad alone demonstrates, part of Project-Alamo was not only to locate the most effective manner in which to determine that swing voters would elect Trump, but to spread misinformation about Clinton, falsely claiming, or at very least claiming without evidence, how the ‘liberal elite’ wish to undermine the norms and values of American society. Such illiberal democratic action was a coordinated attempt at undermining liberal democratic processes so to instil a head of state that would re-calibrate the democratic political system through a populist incarnation of itself.[25] These acts seek to mobilise the electorate with an illiberal lens placed over reality to produce a narrative of fiction professing that the values and norms of the nation are under siege by those amongst the electorate themselves. With this, naturally, came division.

Indeed, the same tactics from 2016 are being wheeled out by Trump’s campaign for re-election, especially as far as the spread of ‘misinformation’ is concerned. In one ad, airing in August 2020, a lone mother sits on a bed holding large cards with the dialogue on the cards pointed towards the camera. The cards read the following:

“I’m a mother of four. Joe Biden worried me. He’s weak. Biden embraced the policies of the far left. Biden will raise taxes. Give amnesty to 11 million illegal immigrants. Biden’s trade deals emptied our factories. I’m afraid to say this out loud…I won’t risk my children’s future with Biden”.[26]

Despite attempts since the 2016 election to limit the spread of misinformation, such as the introduction by Amy Klobuchar of the Honest Ads Act in 2017, the Trump campaign has re-engaged with its populist, illiberal, roots. In this advert not a single claim was referenced, cited, or uttered with some evidence provided to add substance to its own assertions. The question remains, will these same campaign choices from 2016 deliver a pay-off for the Trump campaign in 2020?

Although this is explicitly not the question that I would like to discuss, my purpose here will be to discern where political pressure may be applied by both the Trump and Biden campaigns, in order to replay the outcome of 2016 or to reduce support for the ‘Trumpian moment’ that has immanentized since 2015. This will therefore highlight where the election will be fought, i.e. which particular states and even counties the candidates need to focus their attention on. So as to achieve these ends, the chapter will unpack and discuss in the following order: the electoral process for electing presidents, electoral history prior to the previous presidential election, the 2016 election, and then from this detect which areas will be of interest come November 3rd.

The Electoral College

At the time of writing, according to the latest poll of polls, Biden is a clear 10 percentage points ahead of Trump, polling at roughly 52% and 42% of the electorate respectively.[27] Of course, some may suggest that this is a biased poll, being taken from the BBC. Yet, we can see that no matter which poll of polls we take into account, even those taken by Trump supporting organisations like Fox News, Biden tops Trump in every case, and has done with almost every poll taken this year.[28] For some this may constitute the basis for Deja-vu. Throughout the 2016 election race the opinion polls emphasised a similar and established gap between Trump and Clinton, with Trump never overtaking Clinton in the poll of polls, and with two weeks to the election some 6% between the candidates.[29] This sustained and defined gap in the polls we have experienced before, and much like the 2016 US elections and Brexit referendum in the UK, the polls did not correctly predict the outcomes of these tectonic events. This being said, the opinion polls of the US elections did correctly determine which candidate would receive the popular vote, with Clinton gaining 2% more of the electorate than Trump. So why did Trump become president if he did not receive the majority of votes? For those who already understand this, please feel free to skip ahead, for those who do not the answer is simple – The Electoral College.

Since the early nineteenth century, American voters have not voted directly for the president, but for the members of a national Electoral College who do vote directly for the President. In order to ensure that the individual states constituting the federal union retain their sovereignty, it was decided that it should be delegates of these states that decide who becomes the head of the federal government.[30] Each state sends a certain number of delegates (the amount of congressional districts and senate seats combined for that state – meaning that votes are distributed roughly in proportion to population) to a national Electoral College of 538 delegates. In this fashion, to give an example, California sends 55 delegates to the Electoral college, whereas Wyoming only sends 3. California has 53 congressional districts and two senators (as all states do), making 55, whilst Wyoming has only a single congressional district but also two senators, coming to just 3.

The Electoral College functions with a ‘winner-takes-all’ mechanism. This means that whoever wins the popular vote in each state, even if by a minimal number of votes, that candidate will receive all of the state’s electoral college votes. Therefore, for example, in 2016 Donald Trump won the popular vote in Michigan by just over 10,000 votes (See Figure 5), and as such, received its full 16 electoral college votes. The only exceptions to this are Maine and Nebraska – whereby two electors are assigned to vote with the state-wide popular vote and the remaining electors are allocated based upon the plurality of votes in each congressional district. As in Maine at the previous election, this means that a single state’s electoral college votes could be divided between the two major candidates, but only in Maine or Nebraska which have a small number of electoral college votes at any rate.

So how does the president become the President? Simply put, the President is granted the legitimacy to hold office if they receive 270 electoral college votes or more, holding a majority. This explains how it is that, like in 2016 and 2000, a candidate may secure more votes nationally and win the popular vote but not become president. This of course raises questions about the democratic legitimacy of the whole electoral process, whereby a president may be instated without a majority of the popular vote, with a greater share of the electorate voting for other candidates.

 Equally, the Electoral College favours larger states. If a candidate were to win just seven states – New York, California, Texas, Pennsylvania, Illinois, Florida and Ohio – they would amass 209 electoral college votes, even if they win by only a small margin in each state. Subsequently, this makes the issues of these densely populated areas a higher priority for candidates than those of rural and sparsely populated regions.[31] However, as Trump demonstrated, the electoral college votes (ECVs) of these smaller states can add up, and alongside the security of one or two of the most populated states a candidate can win the election without the popular vote of the nation as a whole.

The key dates to remember concerning the Electoral College are as follows: (a) November 3rd – Election Day. (b) December 8th – The deadline day for resolving electoral disputes. There is a month period set forth in the constitution where electoral results may be legally contested before the Electoral College meets. If Trump or Biden wish to dispute the election results, much like Al Gore in 2000, this month will be the space in which they do so. (c) December 14th – The Electoral College meets in each state and cast their ballots. (d) December 23rd – The ballots from all the states’ delegates must be received by the president of the Senate. (e) January 6th 2021 – The US Congress meets in a joint session to count the ECVs, and lastly, (f) January 20th 2021 – Inauguration day. 

Analysing Past Presidential Elections: 1984-2016

Now understanding the manner in which the President is elected, if we are to attempt to grasp where the 2020 election will be fought, the statistical context of this election must be explored by surveying the results of past presidential elections. The period from which I have chosen to review past presidential election outcomes is from 1984 onwards. I have chosen to only examine presidential elections after the 1984 general election for a twofold purpose.

First of all, it made sense to return to an election far enough in the past that it would allow for trends to be discerned, but also not too far that its context is historically disjointed from our own. Beginning with the election of Bill Clinton to office in 1992, or even with his re-election in 1996, would have begun this exploration with a Democrat president that we know secured the presidency in both elections by carrying a healthy 30 states (Figure 2a). I felt that this was too unlike the context in which we find ourselves – at the end of a Republican’s first term in office. To begin with Bush Jr. in 2004 would have of been too soon, and to begin with Bush Snr., with his 1988 election win, would not have matched the context of a Republican president coming to an end of their first term after taking the presidency from the opposing party. In this sense, the second election fought by the then President Regan in 1984 fulfilled this threefold criteria: (a) Regan was a republican who had taken the administration from the hands of Carter’s Democrats in 1980, (b) his campaign was for re-election, and (c) 1984 is not so recent that its utilisation as the statistical point of departure for this study would disenable the detection of holistic trends, and yet not so far in memory that its incorporation would verge on the anachronistic. Furthermore, it is not uncommon for President Trump to compare himself and his policies to that of Ronald Regan. Thus, in both a rational and a somewhat poetic manner the 1984 election rendered itself as the most fitting point to begin this brief historical review.

Turnout and Carrying States

In any election turnout is key. The greater the turnout the greater one may be able to claim the outcome is democratically legitimated. The reverse holds equally. There are two operationalisations of turnout that I will employ throughout this investigation as a whole. The first concerns turnout as a percentage of Voting Age Population (VAP), i.e. the percentage of the population over the minimal legal voting age that turned out. The second measures turnout as a percentage of the Voting Eligible Population (VEP), i.e. the percentage of the population legally eligible to vote that cast a ballot. Although VAP will be discussed and can be useful in grasping predictive demographic changes, because of its widespread use across the US and that it takes into account the sheer number of ineligible voters (those without full citizenship, felons, etc.), an operational aspect that VAP cannot substitute for, VEP will be discussed as the preferred measure, unless specifically stated otherwise.[32]

Turnout in the US is somewhat unremarkable. Relatively low in comparison to other democratic states, for the past century presidential elections have turned out roughly between 50-65% VAP, with a high of 63.06% VAP as a result of Kennedy’s 1960 election to office.[33] As can be perceived from Figures 1a and 1b below, turnout as a percentage of VAP since 1984 has remained no higher than 58.23%, in 2008, and no lower than 49%, in 1996. Granted that this is still low in comparison to other countries, since 1996 turnout as a percentage of VAP has increased, despite a minor hiccup in 2012, to 55.67% in 2016 – a trend set to continue rising – at least with some optimism at the helm. Interestingly, although presenting different figures with its divergent operationalisation, turnout of US presidential elections as a percentage of VEP mirrors this stipulation that although still sitting between 50-65% VEP, turnout rates for presidential elections are, on the whole, slowly rising (Figure 8b).

Aside from this, there are only a few observable trends that concern partisan voting statistics and their relationship to turnout. As presented in Figures 2a and 2b below, concerning the number of states carried, between 1984 and 2012 partisan pay-offs appear in alternate couplets. In 1984 and 1988, the Republican candidates carried the greater number of states before an alternate inversion of this every other election; here the Democrat candidate (Clinton) carried the greater number of states in 1992 and 1996, before Bush Jr. in 2000 and 2004, and Obama in 2008 and 2012. With the second election of each couplet, the turnout (both VAP and VEP) drops in all cases but that of 2004, at the height of the War on Terror and the ‘securitisation’ of US policy, in which Bush Jr. increased both the number of states carried by one to 31 and turnout (VAP and VEP) by 5% (Figures 1 and 8). Equally, increases in turnout occur at the beginning of each couplet and the instatement of a new partisan administration into office. This was the case in 1992 from 1988, 2000 from 1996, 2008 from 2004, and 2016 from 2012.

This begs two questions. The first is whether or not the 2020 election will see Trump re-elected by carrying more states than Biden, in which case the trend since 1984 will continue, i.e. that parties are elected into office in electoral couplets by carrying the greater number of states in two consecutive elections. The secondary question concerns whether or not turnout will fall or rise. Either: (a) turnout drops and Trump remains president, in which case the increase in turnout in the middle of Bush Jr’s two-terms is an anomaly to the trend; (b) turnout increases and a new partisan administration begins, reaffirming the tendency that the transfer of presidential power to the non-incumbent’s party comes with an increase in turnout, but equally the distinct couplets of terms are broken; (c) turnout increases and Trump remains president by carrying the greater number of states, in which scenario the Trump presidency would match the Republican administration under Bush Jr.; or lastly, (d) Trump looses and turnout decreases – bucking all trends.

Nonetheless, as far as carrying states is concerned, stealing them from across the political cleavage is far easier said than done. Between 1984 and 2016, the popular vote of 14 states[34] has consistently supplied Republican candidates with ECVs, comparing drastically to the Democrat’s 2 (Figure 3). In terms of ECVs these 14 republican states have consistently, and perhaps now somewhat predictably, offered a collective 118 Republican votes at every single presidential election since 1984. This compares sharply to the 13 ECVs offered at consecutive elections from Minnesota and Washington, DC to the Democrats. By the time of Bill Clinton’s election to office in 1992 we saw a number of states turn blue and subsequently remaining so since – some 14 of the most densely populated states[35] providing 196 ECVs to the Democrat candidate at every presidential election following Clinton’s rise to office. Having said this, the Republican Party has held a number of states consistently since the election of Bush Jr. in 2000, adding another 8 states[36] to the list of 14 already consistently voting Republican. The number of ECVs that it has become traditional for the Republican candidate to predictably hold sits at 180. This leaves 12 states[37] that have swung like a pendulum between parties at presidential elections since 1984 – amassing to 156 ECVs.

As far as turnout is concerned, there is one final point that I would like to make abundantly clear. Many often utter the phrase ‘If only more people had turned out, we would have won’. What they intend to say is ‘If only more people had turned out for us, we would have won’, an obvious assertion, but what this does not tacitly do is conflate a general rise in turnout with a rise in popularity for any given party – the initial statement does. In their recently published work ‘The Turnout Myth’, Daron Shaw and John Petrocik sought to expand an analysis into US election results and turnout. On the whole, their aggregated data analysis clearly led them to conclude that: “within the range of turnout variation that we experience in this county [The United States], there is no systematic link between election outcomes and turnout levels”.[38] In this manner, an increase in turnout should only broadly strengthen the democratic legitimacy of the winning candidate – there are no trends to show that broad turnout variations and fluctuations benefit either one of the mainstream political parties.

Partisan Vote Share

What about the popular vote on the national stage? Aside from the 1984, 1988 and 2004 elections, the Republicans have trailed the Democrats in both the number of votes, and as such, their percentage share of the popular vote (Figure 4a). This being said, the lead that the Democrats have held in recent years over the Republicans has been varied. For example, in 1996 Clinton received almost a 9% grander share of the popular vote over Bob Dole, whereas in 2016 this buffer fell to just over 2%, and who can forget the 2000 election, where, although not securing the requisite ECVs to hold office, Al Gore received only 537,179 more votes than Bush Jr. This articulates that although we have grown accustom to the Democrats gaining a larger share of the vote in presidential elections since the 1990s, with 2004 aside, the degree of marginal buffer between the two major parties has been considerably varied.

Nonetheless, between 1984 and 2008 the Democrats received a clear rise in electoral popularity. As Figure 4c illustrates, in this time period there was a clear positive correlation between the passage of time and the number of votes that the Democratic candidate received. Nevertheless, in 2012 and 2016 the Democrats lost votes, decreasing their share of the popular vote across two consecutive elections. The assumption in 2016 was that this would increase, given the popular opposition to Trump. As we now know, this was naïve; perhaps we should learn from this error and practice a quiet scepticism concerning an increase in the Democrat share of the vote for this election. Then again, if one undertakes a regression analysis with the number of votes the Democrats received (y) for each presidential election (x), the trend is unmistakably that the Democrats have become more popular with time, displaying a regression coefficient of ŷ = 1025953.42x – 1998241599 and a coefficient of determination (r2) of 0.91819 – indicating a strong positive correlation and probability of further increase.

As the Democrats have begun to lose votes since the election of Obama in 2008, the Republicans seem to be steadily increasing. By focussing on the percentage of the popular vote that each party has received since 1984, as opposed to sheer number of votes, we bear witness to a different story. Despite beginning the period that I have chosen to focus on with 58.77% of the popular vote, falling to 37.45% in 1992 and rising to 50.73% in 2004 (Figures 4a and 4b), the Republican share of the popular vote displays a small negative association with the passage of time. As the trendline in Figure 4b evidences, the Republican share of the popular vote over time exhibits a regression coefficient of ŷ = -0.18x + 403.792, a corelation coefficient (r) of -0.30639, and an r2 of 0.09387. In simple terms, this means that as time passes the Republican share of the vote is somewhat decreasing, irrespective of the rise in the number of votes that the Republicans have collected since Obama’s election in 2008.

How does this compare to the Democrat’s vote share over the past thirty-six years? As discussed above, by looking at the number of votes alone, there is a strong correlative increase between the number of individuals that vote Democrat and the passage of time. However, by looking at the share of the popular vote that the Democrat candidate has collected since 1984, this correlation is not nearly as strong. Although beginning the period with just 40.56% of the popular vote, with the re-election of Clinton in 1996 and the election of Obama to office in 2008 the Democrats steadily increased their vote share to 52.93%. Since 2008 though we can witness a steady decline in the Democrats’ vote share, decreasing to 51.06% in 2012, and then 48.18% in 2016. As the trendline in Figure 4b illustrates, the Democrats’ share of the popular vote over time reveals a regression coefficient of ŷ = 0.27x – 499.024, an r of 0.77690, and an r2 of 0.60358. This tells us that there is a moderate positive linear association from 1984, irrespective of the Democrats’ decrease in vote share since 2008.

All in all, this data reveals that as time passes the likelihood of the Republican party decreasing its vote share is mirrored by the likelihood of the Democrats’ increase. The two are not a perfect mirror, it must be said. The correlation between the Democrats’ vote share and time displays a far greater positive association of correlation (r2 = 0.60358) than the regression we see of the GOP’s (r2 = -0.30639). Simply put, since 1984 the Democrats’ share of the popular vote is increasing at a greater rate than the GOP’s is decreasing.

In 2016, nevertheless, although the GOP’s vote share declined from 47.2% to 46.08% (Figure 4a), it may comfort the Republicans to know that Donald Trump ascertained over 2 million more votes for the Republican Party than Mitt Romney in 2012. If in 2012, by some strange occurrence, Dr. Emmet Brown were to have appeared in his DMC DeLorean, or Dr. Who in their Tardis, and told us that Donald Trump would ascertain millions more votes for the GOP than the then veteran Republican candidate Mitt Romney, this would have probably been more unbelievable than the prospect of time-travel itself. So, what happened in 2016? Which localities were swayed by Trump to make something that would have been unbelievable in 2012 a reality? This will be the chief concern of the following discourse.

The 2016 Presidential Election

The purpose of this section is simple. It is not to explain or discuss who voted for Trump, as this would warrant a study in itself and of which there are already a number.[39] The function of this section will be to briefly outlay the locality of Trump’s gains in order to ascertain some insight into where the 2020 election will be fought. This discussion will draw from Figure 5 below.

Before any individual states can be examined, it is best that we discuss the national results first. As far as turnout was concerned (Figures 1 and 8a), we saw an increase in both turnout as VAP and VEP by 0.8% and 1.5% respectively from the 2012 Presidential election, seeing over 136.6 million ballots cast. Although Trump received a 2.1% lesser share of the popular vote from Clinton’s 48.18%, the electoral preference of 62.98 million voters willed out and Trump won the presidency with a higher number of ECVs; this happening with the election of Bush Jr. in 2000 but not before that since 1888 with the election of Benjamin Harrison. In 2016 Trump achieved 306 ECVs, carrying 30 states – comparing sharply to the 232 ECVs won by Clinton.

As far as regional geography is concerned, although more sparsely populated, the Southern and rural states of the Mid and North West voted predominantly for Trump, especially those states with a high number of self-identifying evangelicals.[40] Conversely the North East, Western Coast and parts of the South West voted for Clinton. This is not a particularly unusual division between Red and Blue States over the course of the past few decades, it has to be said. Yet this being said there were still some surprises, affording Trump his lead of 74 ECVs over Clinton.

The First of these surprising states that has to be examined is Wisconsin (WI). In 2016, WI gave its 10 ECVs to Trump. There are two points worthy of note concerning WI. The first is that this came as an absolute shock to political scientists and pundits, as until 2016 it was perceived that WI was a firm blue state. Indeed, until 2016, WI had not afforded a Republican candidate its ECVs since Reagan’s re-election in 1984. Therefore, come November 2016 it was a shock to see that WI had voted 47.2% Republican, against the Democrat’s 46.45% – a difference slightly over three-quarters of a single percentage point, just under 23,000 votes of its almost 3 million cast.

Pennsylvania (PA) and Michigan (MI) were a similar surprise. Michigan, which commands a moderate 16 ECVs, and Pennsylvania, with its slightly larger 20 ECVs, had not awarded a Republican candidate their combined 36 ECVs since the election of Bush Sr. in 1988. Of these two, although PA wields the greater number of ECVs, the 2016 result in Michigan was astoundingly unanticipated. In MI, of the total 4,799,284 votes cast 47.5% voted for Trump, whereas 47.27% voted Clinton – a difference of under a quarter of a percent, just a tiny 10,704 votes. Largely, this can be attributed to both an overwhelming and yet underestimated support for Trump from working class voters in previously held Democrat population centres outside of larger cities. In PA, out of just under 6.2 million votes cast, the difference between Trump and Clinton came to just 44,292 votes, where Trump received 48.18% of the popular vote and Clinton 47.46%, a similar distribution of vote share to Michigan. Although the areas that voted Republican did not necessarily change, we can see that rural counties voted in greater concentration than before for Trump. As Figure 6g lays out, in counties such as Erie and Lackawanna in PA, alongside Saginaw and Eaton in MI, the Democrats’ impressive majorities of 2012 dramatically reduced and this saw the Republican share of the vote increase – by almost 11% in Lackawanna, PA, a democratic stronghold.

Two other states of interest that turned Red in 2016 are Ohio (OH) and Florida (FL). With just under 5.5 million votes in Ohio, Trump took the former Democrat state and managed to gain not just a higher vote share against the Democrats, but win an overall majority with 51.69% of the popular vote. This was greater than even Obama’s vote share in 2012 and ultimately reduced the Democrat share of the popular vote from over 50% to 43.56%. Areas of OH on the southern coast of Lake Erie, in the suburbs of Cleveland and Toledo, that had voted marginally in favour of Democrat candidates in the past inverted their voting preference, with the same trend occurring in the more rural working-class areas. Interestingly, these counties that inverted their partisan voting preferences are also some of the most poverty stricken, like Turnbull and Ashtabula for example.[41]

Florida, as expected since the ‘Hanging Chad’ fiasco of 2000, did not disappoint with its fair share of dramatic surprises. Out of a total of 9.4 million votes cast, there was just 1.2% difference between the Republican and Democrat vote shares. Trump, in his state of residence, ascertained 49.02%, whereas Clinton won some 47.82%. Florida has always been known to be a ‘Purple’ swing state, never wholly Democrat nor Republican. In this manner, it is always mildly unpredictable as to which party will pick up its essential 29 ECVs. There are potentially three explanations for the success of Trump in FL. The first, similar to elsewhere, is his appeal to the white-working classes. More interesting and unique to the demographics of FL, rather, concern: (a) the average age, and (b) the number of Cuban Latino voters. Florida has one of the oldest populations in the US, with almost one fifth of its population over the age of 65.[42]

This is significant because we know from a number of studies which have since been undertaken that the over 65s nationally polled better with Trump, where 53% of registered voters over the age of 65 reported voting Republican in comparison to the 44% who voted Democrat.[43] In this manner, the older the population, the greater the probability that the location would have voted for Trump in 2016 – explaining how Trump picked up so many votes in FL. Alongside this, two-thirds of America’s 1.2 million eligible Cuban-Latino voters reside in Florida, and although we can see that, on the whole, Latinos voted in favour of Hillary Clinton (62%), Cuban-Latinos, who are often related to Cuban emigres fleeing the Castro Regime, voted predominantly for Trump (54%).[44] Perhaps we can therefore suggest that Trump held Florida before even a single ballot was cast, with its unique demographics playing to the Republican’s benefit. Although to suggest this would be to ignore the previous elections in which FL voted popularly for the Democrats, winning its 29 ECVs was crucial to the Trump victory.

There are a number of other states that could be discussed at some depth that led to the close results. States such as Arizona, Colorado, Georgia, Minnesota, New Hampshire, North Carolina and Virginia all held margins of difference between the popular vote of Trump and Clinton of less than 6%. Although it only accounts for 4 ECVs, the thinnest margin was by far in New Hampshire. Here, Clinton secured 46.83%, with Trump not far behind at 46.46%. This 0.37% difference amounted to just 2,736 votes, out of the total 744,296 cast. Although Clinton won the 4 ECVs on offer, one would find it difficult to imagine a finer outcome. The same could be said for the result in Minnesota. Minnesota is the only state apart from Washington DC to have voted Democrat consistently in the time period under examination, as stated above. Even here, in 2016 there was only a 1.5% difference in vote share between Trump and Clinton, some 44,756 votes. Much like New Hampshire, although Clinton secured the 10 ECVs on offer, Trump made the race perhaps too close for comfort, especially in a state that has consecutively voted Democrat in high numbers for forty years at presidential elections.

Indeed, although six states metamorphosed red unexpectedly, winning Trump the election, no single state flipped its popular allegiance to the Democrats. Not a single state that voted majority Republican in 2012 voted in majority for the Democrats. This signals that the national electorate favourably responded to Trump’s ‘Republicanism’ in a way that the electorate did not for Clinton. As discussed above, because the election of the president is undertaken by the electoral college, not the people directly, it is not enough to secure widespread support, but one must secure widespread support in the right areas. This brings us to the election bearing down upon us at this moment. What are ‘the right areas’ for each party? Where will the election be fought on the third of November?

Election 2020: Areas to Watch

The purpose of this segment will be to lay out where the Presidential election will be fought, i.e. which states have a high possibility of swinging based on past election results and which states the candidates will be eager to consolidate or recover from the opposition. First of all, these are not scientific predictions. I shall focus my attention on just 6 states, laying out why they are important for the candidates to defend or secure and why we should expect some emphasis to be placed on the issues of these states. Alongside this, I will lay out a handful of key counties in each of the six states that one would expect to be campaigned for in order to secure these states as a whole. Overall, this section will place its emphasis on Wisconsin, Pennsylvania, Florida, Minnesota, Michigan, and North Carolina. These are the six that will decide the election in November, at least as far as the outcome of the 2016 election suggests. Between these six states alone there are 100 ECVs, and each state could swing either way. I will here predominantly focus on Figures 5 and 6.

 There are of course a few other states that are spoken about with some emphasis in the wider media, namely: Georgia, Indiana, Iowa, Ohio and Virginia. I have chosen to exclude these states from my investigation for a handful of reasons. Firstly, although Georgia did give a close result in 2016, with trump only winning by just over 5% of the popular vote, Georgia has not given its ECVs to a Democrat candidate since Bill Clinton in 1992. Such long-term trends are always there to be broken, but should be treated with a sense of reserved scepticism. Although normally swing states, in 2016 neither Iowa nor Indiana were particularly close races, with a 19% lead for Trump in Indiana, and almost 9.5% in Iowa. This is not to say that Iowa could not swing blue once again, but it only holds 6 ECVs, and although Indiana contains 11 ECVs, with that big of a gap there would have to have been a rather serious shift in perspective concerning the approval of the administration for Trump to lose.

 Ohio, on the other hand, is an important state. It commands 18 ECVs and in 2016 there was only 8.13% difference between Trump and Clinton. Conversely, much like Indiana and Iowa, if Biden were to re-secure Ohio it would have to come with an appeal to Trump’s base, and the prospect of this occurring in the Mid-West seems limited. Alongside this, although not wholly comparable, in the 2018 Midterm Election for The House of Representatives, the Republicans still secured 50.9% of the popular vote, implying that Ohio may still form part of the Trumpian heartland (Figure 9b). A similar trend can be specified for Virginia but concerning the Democrats, in which although there was only a 5.32% difference in popular vote in 2016, at the Midterms Virginia strengthened its Democrat hold – and so we should expect it to remain blue.

Wisconsin – 10 Electoral College Votes

Let us begin with Wisconsin. If we look at Figure 6a, we can see the electoral map correlating to the results of the 2016 general election. The first important point to reiterate is just how close this race was. Although Trump won the 10 ECVs, becoming the first Republican to do since Reagan in 1984, there was only 0.76% difference in the popular vote between Trump and Clinton, just 22,748 votes. Why is this significant?

For Trump, the consolidation of Wisconsin would be as symbolic as it would be pragmatic. A state that had held out from voting in favour of a Republican candidate for some 32 years would symbolise, for Trumpians at least, that a demographic of potential Republican voters has been existent in full view for decades, and yet ignored by both parties until Trump’s rise – paralleling the populist notion that Trump speaks for ‘the masses’ that have been neglected. For obvious reasons, Trump cannot afford to gamble with 10 ECVs, and as such, his focus will be to consolidate his gain in Wisconsin. For Biden the reverse stands. Biden will be seeking to regain Wisconsin in a bid to gain needed ECVs, of course, but on a symbolic level the necessity to regain Wisconsin will be part of his nationwide bid to roll back the faith previously placed in the Trump administration; a faith he campaigns was misplaced. Wisconsin giving its popular vote to Biden, after astonishingly turning Red at the previous presidential election, would thus become emblematic of an American people casting Trump’s politics aside – a symbolic turn that the Democrats would not pass up.

If we take a closer look at the electoral map in Figure 6a, we can see that aside from Menominee and the counties that sit on Lake Superior in the north, the highly populated urban areas of Milwaukee in the south east, Madison in the south, Stevens Point (Portage County) and Eau Claire in the centre, and La Crosse in the west, these were the only counties to give a greater share of the popular vote to the Democrats. Equally, a number of well populated counties that had voted in majority for Obama in 2012 voted in greater numbers for Donald Trump than Hillary Clinton in 2016. In reviewing the shift we can observe in Figure 6g, counties such as Crawford, where the Republican share of the vote increased in 2016 by over 10%, or Sauk, where the Democrat vote shrank by over 11%, the Democrat vote decreased hugely, but a decrease that was not mirrored in an increase in vote share for the Republicans.

If we look at all off these counties that were previously Democrat majority holds, the Republicans may have a greater share of the vote, but none a majority or with even so much as an 8% difference. If we focus on those counties that were held by the Democrats, in Green County, where the democrats decreased their vote share from 2012 by some 10%, or in Eau Clair and Portage, the preference of vote share remaining with the Democrats hangs on a knife edge with slim majorities that can be measured in the hundreds of votes. If Trump wishes to consolidate Wisconsin, he will seek to convince the individuals of these counties to vote Republican – achieving his pragmatic and symbolic goals for the state.

Although it is bad methodological practice to compare two different modes of elections, what did the 2018 midterms tell us about Wisconsin? Before I answer this question, I would like to point out that utilising the data from Congressional midterms in comparison to the previous presidential elections is a dangerous game to play. First of all, as Figures 8a and 8b below clearly demonstrate, turnout for midterms is always drastically lower than election years with presidential elections. In conjunction with this, although elected on the same day as the President, the turnout of Congressional elections is consistently lower than the turnout to vote for President. This means that they do not fair as a substitute for comparison in terms of the frequency of voters, much like comparing the outcomes of a large-n study to a small-n study. This is just not methodologically sound.

 Secondly, as the 2016 elections in Wisconsin directly illustrate (Figure 9a), although voting on the same day to staff separate offices of government, the electorate may not popularly vote for the same party across elections. Even though Wisconsin voted with 47.2% of its popular vote to award Trump its 10 ECVs in the Presidential election, the Democrats won the popular vote with 49.75% in the elections for the U.S. House of Representatives. In this way, even comparing election results from the same day in the same state with the same electorate can yield two different sets of results.

 Thirdly, and lastly, elections to the Congressional House of Representatives do not follow county lines, but constituencies are divided into populously proportionate congressional districts, merging and dissecting the counties that are the topic of discussion at the state level of presidential elections. This makes locating how any individual county voted in an election to The House of Representatives an entire investigation in itself.

Despite these three points of caution, what we can do is tentatively compare the results of Congressional House elections in 2016 with that of the midterms in 2018 and use these results cautiously as an indicator of popular support for either party. As just explained, although voting for the Republican candidate at the Presidential election, Wisconsin gave the greater share of its popular vote to the Democrats at the election for the Congressional House of Representatives in 2016, with 49.75% of the popular vote (Figure 9a). At the midterms in 2018, the Democrat share of the vote for the House of Representatives grew to 53.18%, and the Republican share dropped by 0.19 points to 45.61% (Figure 9b). This informs us that support for the Democrats, by at very least some formal electoral and federal concrete measure, over the first half of the Trump administration had increased. The likelihood that the popular vote of the Democrats will increase has become more so, but far from certain.

Although Biden will seek to capitalise on the apparent increase in vote share experienced in 2018, hoping to take back a number of the key counties lost in 2016, Trump will equally attempt to exploit the notion that the Republican vote share in Wisconsin at the midterms did not plummet, implying that key states can be consolidated further to prevent Biden taking back those all-important 10 ECVs.

One excellent example of how Trump intends to do this is in Kenosha County. After the shooting of Jacob Blake by Kenosha Police in August 2020, civil unrest and protests broke out in the city in conjunction with the wider Black Lives Matter movement. Although generally peaceful by day, confrontations with law enforcement became a usual occurrence after dark; this led to the destruction of property and the death of two protestors at the hands of seventeen-year-old Kyle Rittenhouse. Upon visiting the site of the unrest against the wishes of both the Governor and Kenosha Mayor, Trump saw to visit locals who had been affected by the demonstrations. Whilst visiting, his remarks continually related back to ‘Law and Order’ policies, reminding the locals at every moment that the [Democrat] governor and Mayor could have requested federal assistance at an earlier time, but did not. To a Mr. Rode, of ‘Rode’s Camera Shop’, concerning the Governor and Mayor, Trump pronounced:

“Unfortunately, they had a few days when people wouldn’t call us. They didn’t want to have us come in. They just don’t want us to come in, and then destruction is done. A day earlier, we would’ve saved your store…So, the governors have to call, the mayors have to call. As soon as they call, the federal government will come in; it’ll put it out… And it happened in Minneapolis also. Came in — it went for nine days. And we came in; it ended almost from the minute we came in. But these governors don’t want to call. The mayors don’t want to call. And they have to call, and they have to ask.”[45]

Irrespective of whether or not we ethically agree or disagree with Trump’s spinning of an individual’s plight into pitching to a voter, what we can observe is the manner in which Trump will seek to consolidate his hold in areas affected by civic unrest, like Kenosha – by emphasising that Democrat office holders have been ‘weak’ on issues of law and order, leading to the unrest of the times. This is a tactic that has worked with Republican candidates, like Nixon, in the past, cutting across class-distinctions and regions so to display the Democrats as permitting violent crime.[46]

Equally, among all of this, Wisconsin still has questions concerning the level of its voter suppression. Although there is no general agreed-upon definition of ‘voter suppression’, it can be understood as an occurrence that disenfranchises voters through formal structural means that have been deemed lawful, a great example being the requisite of many states that a voter requires a particular kind of Picture ID to vote – which Wisconsin enforces – disenfranchising certain citizens by precluding them the capability to cast a ballot or withdrawing their ballot as void; if this is strategically utilised to discourage a particular demographic from voting, or to limit the political power of a section of society, it comes awfully close to ‘voter dilution’[47].

Returning to the issue at hand, in December 2019 Paul Malloy, a circuit court judge of Ozaukee County, ordered the removal of 230,000 voters from the state registry, claiming that state law compelled him to do so.[48] It was later found that of these 230,000 voters to be removed from the electoral roll, a majority resided in the highly populated areas, of which the Democrats held in 2016 and thus potentially thinning out the Democrat vote share in a purple state to be.[49] Nonetheless, the state supreme court has overturned this purge of the electoral rolls, and all 230,000 have been eligible to re-register since March 2020. The question really is if this will have an effect on the Democrat vote share, i.e. will a number not re-register that would have otherwise voted for Biden?

With all of this in mind, Wisconsin will be an interesting state to watch come November 3rd. Biden will seek to regain those areas that shifted to the Republicans as a number of them were won with only a slim difference in vote share, whilst aiming at appealing to those populated areas that constitute his traditional base. Trump, on the other hand, will seek to make this difficult for Biden by appealing to the same sensibilities that allowed him to turn Wisconsin red for the first time in a generation – his populism.

Pennsylvania – 20 Electoral College Votes

Much like Wisconsin, Pennsylvania (PA) turned red for the first time since the 1980s (Figure 3). The largest difference between them is that PA holds double the number of ECVs. In this manner, alongside Florida, it is vital that the winning candidate will seek to secure the larger share of its popular vote.

In 2016 there was only a slight difference in vote share between Trump (48.18%) and Clinton (47.76%), equalling out to just 44,292 votes (Figure 5). For this election therefore, PA will be considered a ‘Purple’ state, i.e. able to be held by the Republicans or swing back to the Democrats. If we look at the electoral map of the 2016 election in Figure 6b, one can observe a sea of scarlet red, a handful of areas lightly shaded crimson, and only pockets of blue. This indicates that aside from Allentown and Philadelphia in the east, one of the country’s most heavily populated areas, Scranton in the north east, Pittsburgh in the west, and Harrisburg alongside Centre county in the middle of the state, PA voted largely for Trump in all other counties, especially in rural localities. At this election, Trump will simply endeavour to retain these rural majorities. To do anything else would mean appealing to those in demographics that populate Democratic strongholds, which on the whole would more than likely be a futile exercise.

In Figure 6g, we can see that counties such as Erie and Northampton that voted predominantly for Obama in 2012 fell to trump by a margin of just a few percentage points. In Erie county, for example, in 2012 the Democrats received 57.36% of the popular vote. In 2016 this was shaved by almost 11%, with the Republican vote share increasing by roughly 6.6%. These, much like those won with fine majorities in Wisconsin, are areas of the country that have consistently voted Democrat in the past, and are now incredibly close, with only a few thousand votes of difference (Figure 6b).

One of the more interesting qualities we can establish from the 2016 election in PA involves the number of close counties. As Figure 6b displays, all but six of the Democrat held counties returned a vote share that did not exceed a 5% margin of difference to the Republicans – Centre (2.4%), Dauphin (2.9%), Lackawanna (3.5%), Lehigh (4.7%) and Monroe (0.7% – just 532 votes). This is equally mirrored in that only Erie and Northampton counties delivered a Republican win that did not exceed even a minimal 5% margin of difference – with Erie holding a 1.6% and Northampton a 3.8% lead on the Democrats. On a first glance alone therefore, this signals that the November elections in PA will be fought with these counties in mind.

In Biden’s case, assuming that those in Republican majority counties who did not vote for Trump in 2016 will not do so again, he would be able to pick up the 0.72% state-wide difference (44,292 votes) from 2016 by strengthening his hold in these counties that voted for Clinton by only small margins. The six counties that Clinton held with under a 5% margin, listed above and in Figure 6b, contain between 60,000 and over 120,000 votes as a whole in each county. This provides the space for Biden to concentrate his efforts in PA to the issues of those members of the electorate who once voted Democrat but chose to turn to Trump. For example, in 2012, Lackawanna county returned 61,838 votes for the Democrats, over 63%.[50] In 2016, as Figure 6b stands testimony to, the democrats secured 51,983 votes, a loss of almost 10,000 – a quarter of the requisite number of votes that Biden would need to regain and win the 20 ECVs PA has to offer. If he focusses on returning these counties to just their 2012 levels of support, he can retake PA.

 Alongside this, the Republicans won the 2016 popular vote for the US House of Representatives in Pennsylvania with 53.91% against the Democrat’s 45.7% (Figure 9a) – delivering a paralleling outcome to the Presidential election (unlike Wisconsin). At the midterms in 2018 the tables turned. The Democrats received 55.03% of the popular vote, gaining almost 10% of the vote share directly from the Republicans, who ascertained 44.75% (Figure 9b). Perhaps this tells us that the popular vote will once again sit with the Democrats in PA. Although the state may still swing either way, with Trump’s large number of majority held counties, it is the President’s state to lose and Biden’s to win. This will be a make or break state of this election, much in the same way it was in 2016 – although perhaps the Democrats will realise that this time around.

Amidst all of this, something to keep in mind is that the electoral administration in PA has already been accused of potential voter suppression. With such a high expected number of potential mail-in absentee ballots, a number of states, sixteen including PA, will require voters to place their mail-in ballot into a ‘secrecy envelope’ prior to it being mailed. What makes the PA authorities different to other state electoral administrations is that Pennsylvania will become the first state to disqualify ballots if they are not mailed within the secrecy envelope – known as a ‘naked ballot’. Although backed by the PA Supreme Court, and thus with state law on the side of the state electoral administration, it has been estimated that this could disenfranchise some 100,000 voters without recourse to amend the error.[51] In his discussion of this, Richard Hasen, the American UCI professor of electoral law, contends that although the data indicates that mail-in ballots will favour Biden, with more Republicans voting in-person by all likelihoods, throwing out ‘naked ballots’ could come at a cost to the Democrat share of the popular vote in potentially the deciding state of this election – where less than a single percent of the popular vote decided its outcome in 2016.[52]

With a high number of mail-in ballots expected as a result of the pandemic, we must also remember that unlike most other states that are somewhat used to a high number of mail-in ballots due to their size, like California or Texas, the PA electoral administration is not. On top of this, state law does not permit the pre-canvassing of absentee ballots. This means that the suspected uncharacteristic wave of mail-in ballots will not be able to be legally counted until election day, with ballots cast in a polling places taking priority. This contrasts to say Florida, for example, which will begin counting mail-in ballots before November 3rd. This is significant because if the election comes down to the winner of Pennsylvania’s 20 ECVs, which as I have shown above, it may well do, we may have to wait weeks for the result. If this is the case it could lead to a number of wholly negative outcomes, from candidates claiming victory prior to the formal announcement of results, to protests and discord on the streets between voters, with consequentially a mass of prospective electoral litigation on the horizon. Therefore, as far as PA is concerned – watch this space.

Florida – 29 Electoral College Votes

Of the swing states that I am investigating, Florida (FL) is arguably the most significant. It has the highest number of ECVs of the swing states – 29 – and always remains complex to predict as a result of its high and varied population demographics. In this way I should begin this segment by stating that it simply could go either way – as history continually reveals (Figure 3).

In 2016, Florida held just a 1.2% margin of difference between the Republican (49.02%) and Democrat (47.82%) vote shares, indeed only 112,911 votes in one of the most populated states, as Figure 5 displays. Equally, by looking at the congressional election results for the US House of Representatives from 2016 and 2018 (Figures 9a and 9b), although the Democrat’s vote share increased and the GOP’s decreased, there was only a few percentage points of deviation between the two – perhaps suggesting that attitudes towards Trump since he took power had changed very little. As discussed above, it has often been stated that a big part of Trump’s base in Florida – his state of residence – concerned social demographics that FL holds a uniquely high number of, such as Cuban-Latinos and the over 65s. But, this said, which counties should be the focus of intrigue at this election?

By looking at Figure 6c, we can gain a better insight into answering this question. In 2016, the Democrats collected the popular vote in only 9 counties that are highly populated, as is a definite trend across the country. These counties were as follows – Leon and Gadsden counties surrounding Tallahassee in the North; Alachua county (Gainesville), Hillsborough county (Tampa), and Orange and Osceola counties (Orlando and Kissimmee) in the centre; and lastly, the densely populated south eastern coast-line that contains Palm Beach, Broward and Miami-Dade counties. All else returned a Republican majority, even if they were slim. The slimmest majorities of these were four Republican counties (figure 6c): Duval (1.38% difference), Pinellas (1.12% – less than 6,000 votes), Seminole (1.57%), and St. Lucie (2.42%). Not one of these counties should really be penned as ‘sparsely’ populated either, with tens of thousands of voters turning out. How have these counties changed over time however? Have they always been this close?

Let’s begin with Duval county. Interestingly, both parties reduced their vote share in Duval county from the Obama-Romney election in 2012 to the Trump-Clinton race of 2016. If we take a look at Figure 6g, we can observe that although the Republican’s won the popular vote of the county, they lost 2.35% of the vote share from 2012. Contrasting this is the Democrat popular vote, in which although they too lost support, it was a fraction of the size of the loss the Republicans faced, at only 0.13% of the vote share. Essentially, this conveys that the vote share lost by the Republican’s in the county, if harnessed by Biden in November, would turn it blue. This discloses that there is hope for Biden in Florida, but he would need to campaign to regain those voters disillusioned with the two major parties to the extent that they either voted for a smaller party (Greens, Libertarians, etc.), or not at all. This will be his task throughout Florida if he seeks to win the state – to appeal to those on the periphery of both parties; Trump will be attempting to do the same, but defensively.

Pinellas and Seminole counties could not have been more different in the outcome they returned in 2016, besides their mutual support for Trump of course. Pinellas was previously a Democrat county, holding 52% of the popular vote in 2012. Then again, in 2016 the Democrats lost 5% and the Republicans picked up some of these voters, all be it not gaining any substantial amount of the vote share. This too was the outcome for St. Lucie county. Pinellas is typical of the majority of counties across the Trump-won states in this respect. Nonetheless, there is a substantial difference with this and the change we can witness in Seminole since 2012 – a result that is wholly atypical. Although remaining Republican from 2012, when the GOP received 52.57% of the vote share, in 2016 Trump lost almost 4% of the popular vote in Seminole, and consequently the Democrats gained a tiny amount – just 0.97%. Seminole is therefore the only county I will discuss where the Republicans won in 2016 but still decreased their vote share in comparison to the Democrats’ improvement on the Obama years. This is the reverse of what was typical in 2016. Why is this important, it could just be anomalous?

This is true, the results in Seminole may just simply be an anomaly. Despite this possibility, these results inform us that there is a demographic in the largest swing state that were uninspired by the Trump campaign to the point at which the Democrats could pick votes up, no matter how small a percentage. This discloses to us that Florida will have to be defended by Trump to a different extent than other states, not only because he secured them in the previous election, but because in the last election there were already murmurs beneath the surface that showed there was always a possibility that Trump could lose votes for the GOP. Biden needs to take advantage of whoever this demographic is if he wishes to win Florida. Assuming that Trump’s character alone will be enough to win was the folly that Clinton made here in 2016. This should not be made again because every vote in Florida counts. 1992 aside, the candidate that secures Florida wins the election as a whole (Figure 3). This summates neatly the importance of the Sunshine State.

Minnesota – 10 Electoral College Votes

Minnesota returned an interesting result in 2016 simply because it was an exceptionally close race between Clinton and Trump. I have chosen to include it in this list of interesting swing states to focus one’s attention on come November third for a number of reasons. The most significant of these reasons concerns the fact that all the other swing states that I focus my attention on here handed their ECVs to Trump in the last election. I am implying with this utterance that these are states Trump might lose the support of if he cannot effectively defend his 2016 gains. I wanted to include at least one state that was held by the Democrats in the last election that possessed the potential to turn red, at least as far as the electoral data is concerned – a state that although has consistently voted for the democrat candidate in the past, a small voice in the back of Biden’s mind might be raising some subtle concern about. This state is Minnesota.

As we can see from Figure 3, Minnesota has consistently voted in favour of the Democrat candidate since 1984. Indeed, the last election that Minnesota gave the greater share of the popular vote to a Republican candidate was with Nixon’s re-election in 1972. When vast swathes of the population were voting for both Reagan and the Bush duo, Minnesota was the only populous state, excluding Washington D.C, to consistently tread its own path.

 In 2016, the popular vote divided the state without either party receiving an absolute majority. From Figure 5, one can distinguish that Clinton received the greater number of votes, with 46.44%, and Trump achieved 44.92%, with only 44,756 votes separating the outcome in a state where almost 3 million voted. In 2012, Obama had beaten Romney in MN by some 226,000 votes, gaining 52.65% of the vote share, with Romney winning only 44.96%; the usual spread and distribution of votes between parties that could have of been expected from a presidential election in Minnesota.[53] The only other time in living memory where the results were so close was the 2000 election between Bush jr. and Al Gore, in which Gore took 47.9% and Bush 45.5%; this makes it closer than 2016 in terms of difference in vote share, but this equated to 58,607 thousand votes, close to 14,000 more than in 2016.[54] In this manner, I think it would not be a dramatic but somewhat realistic exercise to claim that 2016 was potentially the closest contemporary presidential election we have seen in Minnesota. Does this guarantee that it will be a swing state in 2020?

In simple terms, no, absolutely not. As I have already stated, Minnesota has consistently awarded its 10 ECVs to the Democratic candidate throughout the period in which I have focussed this study. If it is a somewhat usual cliché to hear that most the southern states will always vote Republican, the same stands for Minnesota – this is the status-quo. What is significant is just how close the previous election results were. The only other election where it was closer in terms of difference between the popular vote of the two major parties was in 2000, and this was the very phenomenon that made this particular election famous nation-wide. Thus, with this in mind, what separates the 2016 election results in MN to that of 2000 is that it was an outlier in comparison to the other Democrat held states.

For Trump then, although it would be an uphill battle against precedent, a victory for him in MN would be much like the scenario in Wisconsin – both a pragmatic and symbolic victory. If Trump were able to take Minnesota, if even by just a single vote, it would be the clearest electoral legitimation and mass-vindication of his policies that we would have yet seen. Although perhaps unlikely, if Trump really wished to focus on Minnesota in order to secure a victory, it would signal the clearest of indications that the Trumpian moment in US politics has ossified into something greater than a mere moment, but become an era.  

Biden will be aiming to disrupt this potentiality for Trump in any way he can. If Minnesota turns red, Democrat partisan politics as we know it would change. This is not an outcome that Biden would wish for, to any degree, as it would lead to a cast iron division in the party between those who would argue that some compromise is needed with the right so to re-appeal to the core Democrat voter – that Minnesotans have come to epitomise – for the sake of electoral popularity, and those who would take a more principled stance against the onset of the Trumpian right. For Biden, losing Minnesota would be just as symbolic and consequential as it is for Trump if he were to steal it from the Democrats, the difference being that if Trump does not achieve his goal then there would be no change on 2016. This makes Minnesota an interesting state to watch throughout this election – for its symbolic value more than anything else.

Alongside this, we should not forget that Minneapolis was the epicentre of the killing of George Floyd and the consequent wave of demonstrations over the issue of Race in the US during the course of the summer. Irrespective of whether or not this wave of political action is conceptualised in line with ‘riots’, ‘criminality’, ‘protest’, ‘resistance’, ‘liberation’, and so on, it would not change the fact that such action illustrates a discontentedness with the status-quo in Minnesota. Given Trump’s rhetoric and response to the Black Lives Matter protests and related demonstrations only serving to make these events increasingly conflictual, such as forcefully removing a crowd in Washington, DC to take a photo opportunity in front of a Church whilst brandishing a Bible, I thoroughly doubt that such action will aid his electoral prospects in Minnesota. This being said, stranger things have happened in Presidential elections.

Nonetheless, which counties will Biden and Trump be aiming at appealing to the major demographics of in November? If we review Figures 6d and 6g we can get an idea at the close counties from 2016 that will be interesting to observe in the upcoming election. In Figure 6d, the electoral map of MN from 2016 clearly shows where the wider geographical division rests between Republican and Democrat voters. The Democrats secured Minneapolis and St. Paul in the south east, the counties on Lake Superior in the North, and Rochester in Olmsted county like an island of light blue in the south; all else is a shade of red. For Biden, the demographic voter in areas like Olmsted is where he needs to re-appeal to the electorate. In 2012, every single county that Clinton won in 2016 by the skin of her teeth were largely Democratic-majority areas. In Washington county, for example, which is highly populous and hence more than likely to contain within it the Democrats’ voter base, the Democratic share of the popular vote fell by almost some 3% (Figure 6g). This may seem trivial, but in a state where every vote does indeed matter, handing the Republicans directly 3,000 votes is a blunder.

Undeniably, where Biden will need to focus his attention in Minnesota are areas like Washington county. Sitting just outside Minneapolis, Washington county is a highly populous area that seemed to wholly lose faith with both the major parties in 2016, where the Republicans lost 4% of the popular vote and the Democrats 3% (Figure 6g), even though they still held control of the county. This reveals that there is some 7% of the county’s electorate who Biden can try to either (a) win back, or (b) convince that he is the better of two evils. The same can be said for Dakota county, which is higher in population than Washington county, where the Republicans lost 5% of their popular vote from 2012 (Figure 6g) – these are areas that were clearly not enamoured with either Hillary or Trump, but he might be able to convince them that he is different to both.

Equally, if Biden is serious about holding and consolidating Minnesota as Democrat heartland, he needs to retake votes in areas that swung red. Take Winona county for example, a medium sized county in which the Democrats once held an absolute majority. From 2012 to 2016, Clinton lost the Democrats almost 12% of the popular vote from 55.01%, whilst granting the space for Trump to therefore gain just a few thousand votes to win the county’s larger vote share. These should be areas Biden will focus on to regain those traditional blue votes lost four years ago.

Simply put, for Trump, if he chooses to go for Minnesota he only has to gain 44,756 votes. This means appealing to those in counties that he has already taken from the Democrats in greater numbers. Trump won these counties unexpectedly in the last election. If he were to campaign with their issues at the forefront of his advertising and public meetings, he might stand a chance at not only retaining the votes of those taken by the Democrats by default of their loss, but convincing the electorate positively that he is a candidate worth legitimating the programme of.

All of this being said, the probability of a Democrat hold does sit with Biden. This we can see with the increase in popularity the Democrat’s received in the 2018 midterms, mirrored by the loss of the Republican vote share. In 2016, the election to the House of Representatives returned a Democrat vote share of 50.15%; by 2018 this had increased to 55.13% and the Republicans won 43.68%, a 2.98% reduction from 2016 with a high turnout state-wide for a midterm (Figures 9a and 9b). This points to an increase in popularity for the Democrats, and so this might constitute some evidence to suggest that Biden will win Minnesota. Nonetheless, the Minnesota results were too close for comfort if you were a Democrat in 2016, and this was unexpected. Maybe the unexpected can happen once again.

Michigan – 16 Electoral College Votes

To some extent Michigan was discussed in the above examination of the 2016 election. Just 10,704 votes decided the placement of Michigan’s 16 ECVs with the Trump campaign, some 47.5% of the popular vote for Trump and 47.27% for Clinton (Figure 5). Indeed, with both large rural and urban population distributions, the stakes are high in Michigan. By reviewing Figure 6e we can observe that outside of population centres in and around Detroit, Marquette and Muskegon, Trump won the popular vote to some degree. This, like WI and PA, was largely unexpected as Michigan had not given its ECVs to a Republican candidate since the election of Bush Sr. in 1988.

Trump secured the rural areas of Michigan, as examined above, by appealing to the working-class rural population of the state who felt simply left behind by Washington DC. In this, my estimation is that Trump will seek to consolidate his win but by the same extent he did so in the previous presidential election. In the centre of the state a number of counties  turned red, but only by a hair’s bredth. Although Trump could win the November election without MI, 16 ECVs is not a number that Trump can refuse to pay attention to, he must try and hold MI if he really does seek to secure the presidency and consolidate his programme of political action broadly. This he can do by aiming at re-running his formula for 2016 in appealing to his base, and following the trends that were established some four years ago in the more central counties.

For example, let us look at Saginaw county. In the centre of the state, north of Detroit and south of Lake Huron, Saginaw county holds both populous and rural geographical areas, almost acting as smaller model of the state as a whole. In 2012, the popular vote of the county was won by the Democrats with 55.46% of vote share, trailed by the Republicans with 43.56%, a clear difference (Figure 6g). Come the election in 2016, Trump somehow managed to reduce the Democrat vote share by a staggering 8.39% (almost 10,000 votes), whilst increasing his own party’s by 4.65 points to 48.21% and becoming the more popular party of the area (Figure 6g). This swing is the key to Trump’s second victory in Michigan as it clearly indicated that a sizable percentage of the electorate did not wish to legitimate a Democrat head of state. If Trump can tap into discerning why such a large portion of the electorate here chose to swing Republican and apply it to the entirety of the state, his consolidation of Michigan may be easier than many think.

As for Biden, his task is a simple one – the same as it is for WI and PA. Biden will need to regain the areas where the Democrats haemorrhaged votes in the last election. If I turn our attention back to just Saginaw, if Biden were to focus on returning this county to its previous status as county where the Democrat vote share enjoys an absolute majority, alongside a few votes in other counties like Genessee and Oakland where the Democrat share in the popular vote took a steep downturn (Figure 6g) the state will be his.

If we look at Figures 9a and 9b, laying out the elections to the Congressional House of Representatives in 2016 and 2018, there are indications, as seems to be becoming a recurring phenomenon, that Trump is losing popularity en-masse. From 2016 to 2018, the Democratic vote share jumped from 46.97%, roughly around the same vote share for Clinton as president, to 52.35%. This is twinned with the loss in popularity the Republicans experienced, reducing their vote share from 48.03% in 2016, also roughly close to the state-wide popular vote share awarded to Donald Trump, to 44.61%, a loss of 3.42%. This may seem small, but if the similarity between the Presidential and congressional elections of 2016 continue to mirror one another through 2018 and into 2020, Biden would not only have won back the state but secured an absolute majority.

Although it seems likely that Biden will take back Michigan, it was particularly unprecedented that Trump would win it in the first instance. In 2016, alongside a whole helping of arrogance that the state would naturally remain blue, Clinton did not organise to meet with some of the state’s poorest and precarious workers through arguably one of the most significant labour unions in Michigan – the United Auto Workers labour union – in order to convince them that she had the working classes best interests at heart.[55] Because of Trump’s rhetoric, his character, and his appeal to population demographics that Michigan is bursting at the seams with, if Biden does not mobilise his campaigning capability to the uttermost efficacy in MI, do not by any means be surprised if Trump were to hold it despite the uptick in electoral popularity the Democrats ascertained in 2018. If Biden does win Michigan, this indicates that the Democrats would have re-connected to their base, and thus expect the same phenomenal event to have happened across the country.

North Carolina – 15 Electoral College Votes

Much like the other states that constitute parts of the old southern confederation, North Carolina (NC) has repeatedly voted for a Republican head of state. Throughout the time period that this investigation has made its focus, since 1984, NC has only awarded its 15 ECVs to a Democrat candidate once, to Obama in 2008 (Figure 3). Many consider NC to be a potential swing state in the 2020 election because since 2008 its usual voting pattern, of offering an absolute majority to the Republican candidate at almost every Presidential election since 1984, has been broken.

In 2008, when NC voted majority Democrat for the first time since Jimmy Carter in 1976, Obama won the ECVs on offer with just a 0.32% greater vote share than the Republicans, a shockingly small 14,177 votes difference with both parties returning above 49.3% of the popular vote.[56] From this moment onwards, NC has returned to being a red state, but only by a limited magnitude. In 2012, Mitt Romney won the states’ popular vote, adding to McCain’s returns in 2008. From 49.38% in 2008, the Republicans scraped a small victory in 2012 by increasing this to 50.39% and reducing the number of voters for Obama to 48.35% – this increasing the fine margin of difference of 2008 to 2.04%, a well-won buffer.[57] Trump only increased this gap between support for the Democrats and that for the Republicans. He lost the absolute majority, no matter how tiny it was, that Romney had secured for the party in NC four years earlier. In 2016, the Republicans still held the popular vote despite losing their popular majority, at 49.83%, but the widening of the difference in share of the popular vote between the two major parties came as a result of a collapse on the side of the Democrats’, who failed to secure their 2012 vote share, and declined to 46.17% (Figure 5).

These descriptive statistics are important because what they reveal to us is that a state with a once dependable Republican majority, since 2008, has become unpredictable. Between 2012 and 2016, what made the outcome of the election in NC interesting was that both parties reduced their share of the popular vote, not only losing the Republicans their absolute majority, but returning the Democrats to their pre-2008 level of electoral support. Why is this significant for this election? North Carolina will be a significant state to watch in this election precisely because both parties reduced their percentage share of the popular vote in 2016. This implies that there is sizeable number of voters who can be won back by both parties if their campaigns are effective enough to appeal to the previously disillusioned. For many, this makes North Carolina a potential swing state. Therefore, where in North Carolina will the parties seek to return a higher number of votes than in 2016?

If we take a look at the electoral map in Figure 6f, we see a far wider variation in Democrat and Republican support than in most other potential swing states. Around Raleigh, Durham, and Greensboro in the centre of the state, Charlotte, Asheville and Fayetteville in the south, and a collection of seven counties in the north surrounding Halifax county, the Democrats won the popular vote in a number of localities that were both urban (as in the centre and the southern counties) and rural (as in their northern strongholds). What I find interesting about North Carolina is this region of rural non-metropolitan counties that went against the current in 2016 and voted predominantly for Clinton as opposed to Trump, unlike most other rural non-metropolitan counties in southern and mid-western states.

 By reviewing data provided by the U.S. Census Bureau, these counties do have something that distinguishes them from other rural and non-metropolitan counties in the state – race. These counties are majority populated by Black or African Americans in a state where this demographic makes up only 22% of the population.[58] Alongside this, keeping NC’s northern counties in mind, according to a study undertaken by the Pew Research Centre in 2018, only a mere 6% of Black or African American validated voters reported voting for Trump nationally.[59] In these counties bordering Virginia, a voting block rather crudely known by some as the ‘Black Belt’, in 2016 the majority Black or African American electorate chose to vote with racial issues in mind, as opposed to the populist rural-urban, conservative-liberal divide we know that worked for Trump in other counties in all similarity bar this. This tells us that these counties are more than likely to vote the same way again come November, and therefore Biden already begins the race in NC with a foothold in the popular vote.

The rest of the county voted for Trump in 2016, especially the predominantly white rural counties of the west, that if we review in Figure 6f, voted in high concentrations for Trump – a literal sea of scarlet red. This being said, although Trump clearly commands the popular vote in the west of the county, there is a corridor of mild Trump support in the south between Greenville and South Carolina that only swings Republican, as opposed to voting in the same concentrations for Trump as the western counties did. On the other hand, Trump managed to take some of the states in the north on the border with Virginia that are, as explained above, traditionally Democrat strongholds.

 Take Granville county, for example. In figure 6f, we can establish that during the Obama- Romney election, the county voted 51.75% in favour of re-electing Obama, a majority that fell by almost 4% in 2016 whilst returning a 2.4% rise in vote share Trump at the same time. This led to Trump’s victory in those counties that were just unexpected for him to win. In areas like Pasquotank and Martin Counties, Trump seriously damaged the Democrat share of the popular vote, whilst simultaneously adding to his own. If he can discern exactly what he did that caused this shift in the North of NC and apply it to the locality broadly, who knows, he could increase his vote share in NC as he obviously struck a chord somehow with those who are not the Republican’s traditional voter base.

These are the counties that Biden should focus his attention to if he does was to turn the tables and take a state from the Republicans that, 2008 aside, has traditionally voted for the Republican candidate. A really good illustration apart from Granville county is Nash county. Nash county sits in the north and although rural holds a high population of over 40,000 voters. In 2016 there was only 0.17% of the electorate that separated the vote share of the two parties – less than just one-hundred votes. This is a definition of a divided county, but what makes it different is that it has historically voted in higher numbers for the Democratic Candidate (Figure 6g). Biden needs to retake these areas if he wishes to turn the tide in NC. In my perspective this will be far more complex than it seems, as those areas in west NC that voted predominantly Trump one should not expect to change that much. If Biden seeks NC he will be focussing on taking votes back in the north and picking up votes in the southern counties where Trump only gained a mild majority in 2016. If he can appeal here by providing an alternative that the local electorate deem electable, who knows, it could be another 2008.

Nonetheless, the results from the 2018 midterms still show a republican held state. In 2016, in the elections for the US House of Representatives, 53.22% of voters chose the Republican candidate on the ballot, in comparison to the Democrats’ 48.3% (Figure 9a). If we review how this changed in the 2018 (Figure 9b), NC followed a similar trend to the majority of the rest of the country, that the Democrat vote share increased (here by 1.7%) and that of the President’s party decreased (here by 2.73%) – a normal trend of mid-terms to be perfectly frank (see Figure 7d). This being said, the Republicans still held the majority of the popular vote, with 50.49% of the vote share, indicating to the observer that although some have become disillusioned with the Trump administration, and their Republican Congressional representatives, the state still rests in their favour.

To illustrate this point, NC is one of the few states that has held special congressional elections between 2018 and 2020, with its 9th congressional district holding a special election in September 2019. The 9th district covers both Republican and Democrat held counties from 2016 on the border with South Carolina. Before the results can be discussed, it must be said that the special election took place after redistricting in 2017 and allegations of voter fraud in 2018. In comparison to the election in 2016, there was interestingly no change on the county level as to which party won the popular vote, as it would be bad methodological practice to compare the number of votes. This was the case in all but Robeson county, which voted 50.5% for the Democrats – a swing from the majority vote of less than a percent they afforded to Trump in 2016.[60] This county is just one of many that just ever so slightly voted for Trump more than Clinton, and so it may be telling that just over a year ago it swung blue. What makes this result all the more relevant is that just days after his last debate with Biden, Trump announced a visit to Robeson county before the election, mildly vindicating my own location of these counties as important to Trump and the possibility that these counties have already swung to the Democrats.

Equally, following changes in electoral law concerning Voter ID, North Carolina is now one of the few states in the southeast without restrictions of this kind, an exercise it has yet to experience.[61] Although Democrats locally suggested that the Voter ID laws had an effect on the turnout of African American voters, a number of scholars suggest that Voter ID laws do not have an effect on turnout in any particular mode; either way it will be interesting to see if there are effects of this decision that may favour any one candidate.[62]

My instinct is to suggest that Trump will win NC, due to historical precedent. This being said, if Biden can appeal some way to the average rural white working-class voter in the west of the state, whilst simultaneously appealing to the rural black or African American voter in the north, this could be a close race with significant consequences.

In Conclusion

This investigation has thus far sought to locate where the upcoming November Presidential Elections in 2020 will be fought, and where we should expect Biden and Trump to focus their efforts. If we can say that there was a thesis at all of this chapter, it would have been that this election will be fought in the following swing states: Wisconsin, Pennsylvania, Florida, Minnesota, Michigan and North Carolina. These will be the states to watch as the election creeps closer and closer – especially Pennsylvania.

In order to arrive at this, I chose to first explain the nature of the Electoral College. Although to some this may be an irrelevant or purely descriptive exercise of institutional process, if this electoral mechanism is not wholly grasped than the status of these states as being important, as being potential ‘king-makers’, is lost. As I stated above, it is all well and good placing an emphasis on the national polls in the days running up to the election, but one must always remember that the presidential election is not a nationwide popularity contest – it is a popularity contest within each state, and the decisions of some, e.g. California or Texas, can make or break an election. The hyper-decentralised electoral system that the American political edifice mechanises was described and its most common critiques were laid out, i.e. that it benefits certain states and that the winner of the popular vote does not necessarily become president like in 2000 or 2016. From here, this permitted a review of past elections.

I chose to evaluate the historical context of the 2020 election by trailing back to the 1984 election and reviewing the statistical trends across time that we see in voting data, such as (a) the number of states carried by each party across elections, (b) how states had awarded their respective Electoral College Votes, (c) the popular vote awarded to the two major parties in this period, (d) the number of votes awarded to the two major parties in this period, and (e) historical measures of turnout. On this last point, it was declared that turnout in terms of both V.A.P and V.E.P were rising slowly across the last 36 years. As this investigation is being written, it has been reported that some 50 million Americans have already voted with well over a week left until Election day through early voting, estimating a turnout of 150 million all in all and as such the highest rate of turnout since 1908.[63] The most significant stipulation made here, from my own perspective, came as a result of undertaking a brief regression analysis on each party’s share of the popular vote in Presidential elections since 1984. Broadly speaking, here I projected that a macro-trend of the period seems to be that the Democrats have increased their share of the popular vote of  since 1984 at a quicker rate of growth than the Republicans have decreased their share of the popular vote in the same period – even though both party’s share of the popular vote has been declining since 2012. This is significant as it would seem to suggest that if this trend were to continue in this magnitude, we are more likely to observe a rise in the percentage of ballots voting Democrat, potentially pointing to a Biden win. However, nothing is certain, especially after the 2016 election.

The previous election then became the subject of discussion. Simply put, the extent to which this was one of the closest elections of the contemporary era befell discourse by an examination of the electoral data. Here, I presented to a brief explanation of not how specifically, but where Trump did exceedingly well and Clinton not so well, that lead to the outcome returned. This almost wholly centred on those states that were a shock to have turned red, making them prime targets for assessment as candidates for swing states in the November election. This ultimately led to the discussion in the final, long, section of this chapter, in which each of the six key states detected to be highly probably ‘purple’ were placed a little further under the microscope.

I would like to bring this chapter on the Presidential Elections to a close by reminding the reader that all elections in which an incumbent is potentially re-elected are referendums on not just the kind of President the incumbent is, but the kind of country the President has made. At the beginning of this whole chapter, it was discussed the extent to which Donald Trump was not just a right-wing populist, but a member of the ‘Alt-Right’. Although I can wholeheartedly stand by my earlier stipulation that to claim this in the affirmative is a conflation of phenomena, it will be such a question of extremity at play behind this election. The American electorate will not be going to the polls with Trump in mind in the way they have done with any other politician in the past. With his continued claims that he, the President, is not a politician, he might just get what he wishes for. The referendum that this election is, held concerning his premiership, may not be voted on with rationality in mind – i.e. the success and efficacy of policy – but with the kind of confrontational emotive reasoning that secured him the White House in the first instance. Trump wants to claim that he is not a politician and yet convince the electorate of his successes like any other politician – he cannot have it both ways. Perhaps the irony here matches my four potential outcomes for this election.

1) Biden Wins with the kind of support that the polls are suggesting. The first potential outcome is that Biden will win with such a large share of the popular vote, an absolute majority and more if the current polls are correct, that one should expect Biden to secure 400 ECVs, or perhaps even more.

2) Biden wins with a higher share of the popular vote than Trump, as historical trends have set a precedent for, but not necessarily in all of the key states. For example, if this election were to return the same result as 2016 in all but Iowa, Michigan, Pennsylvania and Wisconsin falling into the hands of the Democrats, this would produce an outcome of 302 – 236 with Biden winning, a similar spread to 2016 but with party roles reversed.

3) Trump wins by securing the same states as in 2016. As far as my own predictive horizon scanning is concerned, I do not see any more states willing to give ECVs to Trump. If this does occur than there has been a serious overlooking of some variable by pundits and political scientists alike that would lead to any demographic of US society awarding some new allegiance to Trump and his Republican party. In this scenario I would expect a similar division of ECVs to 2016 but with the Democrat’s vote share increasing still, perhaps even with 4 or 5 million between the candidates, but only in states that the Democrats already held, like California, New Jersey or Nevada – effectively leading to yet another election where the Democrats win the popular vote but not the Electoral College.

4) Neither wins. There is always the prospect with such unpredictable elections that neither candidate will win. There is of course a slim percentage of the modelling that returns a 269-269 outcome, but it is highly unlikely. This being said, if it does occur, then the 12th Amendment to the constitution would come into effect and the House of Representatives would undertake the process of electing the new President by voting in state groupings, and the Senate would elect the Vice President.

With each of these outcomes, all a possibility, Trump will have to face up to some form of irony. In the first and second cases, he would have lost because of his character, more than likely, unable to convince the electorate of his successes as a politician, despite his efforts to not be seen as one. In the case that he wins, the irony will be in that the electorate would have voted for him on the basis of his character and not necessarily his efficacy as a politician, which he is asking them to do currently. Lastly, if we end up with a tie, Trump is wholly castrated of capability to claim that the election was either won or fully ‘rigged’. The irony is undeniable.

I would like to remind the reader that the candidates have up until the 8th of December in which to contest any results, be them nationally or locally. If it all comes down to a state like Pennsylvania, that is expected to perhaps take weeks to announce the formal result due to the influx of mail-in ballots, the time for litigation will be reduced to just a couple of short weeks. Many contend that in this time-period, Trump or Biden will just declare victory, all be it undemocratically. Alongside this, Trump has already questioned whether or not he will accept the result if he were to lose, and a number of terrorism scholars, such as the renowned academic Bruce Hoffman, have raised serious concerns that the extreme-right will undertake violent, non-monolithic and uncoordinated action in order to coerce a Trump victory.[64] In the case that both of these scenarios occur, or either of them to be honest, it is the average American who loses this election; not Biden; not Trump; the average person will witness their democracy descend into the unknown. All of this makes the November election a political competition that will have shockwaves, not only inside the American political system, but beyond it.

Equally, some may argue that legislative power really rests with Congress and that congressional elections are the true determinate of how the balance of governmental and legislative power evolves in the American political edifice. This will be the thematic kernel of the discourse in the following chapter.

Chapter II.


The first part of this investigation centred its attention on the race for President, which, naturally, has taken much of the attention of the media – considering this is an election for the head of potentially the most powerful state atop the structure of international society, this can hardly be begrudged. Saying this, we must not forget that there are other federal elections taking place on November 3rd, the outcome of which could fundamentally adapt the shape of the American political. At the time of writing, Donald Trump (R) is President of the US, the Republicans hold a 4-seat majority in the Senate, and the Democrats populate the House of Representatives with a majority of 36 seats. I have discussed where is important to shift one’s gaze as far as the Presidential elections are concerned, now I will focus on the upcoming federal elections relating to the legislative branch of the US political system – Congress.

The Basics of Congress

Congress operates on a bicameral basis. This means that there are two chambers of the legislature: (a) a lower house, where legislation is typically proposed, debated and voted on, and (b) an upper house, where potential legislation from the lower house is scrutinised, debated, voted on and passed so the head of state may grant it assent into the political unit’s body of law. In the United States, The House of Representatives is the lower house, comprising of 435 seats, and The Senate is the upper house, constituted of 100 senators – 2 from each state.

Every one of these seats is elected and therefore Congress claims full democratic legitimacy as a logical truth. This indeed makes the separation of powers discussed in the introduction increasingly a self-limiting mechanism of democratic governance because a divided government may legitimately lead to the stagnation of the legislative agenda. As in the case of many presidents, and of Trump currently, an unstoppable force with democratic legitimacy may have its wishes blocked by an immovable object equally legitimate democratically. This gives us just a small insight into the importance of the Congressional election in November 2020.

Before the structure of this chapter will be laid out, and my brief exploration of Congressional Elections commences, how do Congressional elections differ to presidential elections? The first part of answering this question simply concerns processes. The president serves terms of four years, and cannot be constitutionally elected to hold presidential power for more than two terms.[65] Those who are elected to seats in the legislature hold different term lengths depending on the chamber they are elected to. In the Senate, seats are divided into approximate thirds. One-third is elected every other year, determining that Senators serve in six-year terms. The House of Representatives, however, holds a full election of the chamber every other year. This therefore means that in the space of a President’s single term, the midterm elections could see: (a) a shift in up to one-third of the seats in The Senate and (b) the seat distribution of The House of Representatives turn on its head.

One of the greatest critiques of term-length for Congressional representatives came from David Mayhew. In his long essay ‘Congress: The Electoral Connection’, Mayhew discusses the extent to which the representative-constituency link between congressional representatives and constituents is determined by these term lengths. As representatives of the house are up for re-election every other year, they never leave the campaign trail; even after winning another term the potential of losing one’s position is always somewhere on the horizon, just 24 months in the distance. Alongside this, Senators are afforded a healthy six years in which to make changes for their state-wide constituents. In both cases, if a representative does not deliver results with the time they have, they will be elected out of office. Mayhew argued that this has led to a milieu in which congressional representatives are ‘Single-Minded Seekers of Re-election’, where a certain self-interestedness to secure good electoral prospects occupies the rationale behind every action that a congressional representative makes; seeking to lay a claim to any legislative activity, pay-offs, initiatives, policies, and especially funds towards popular projects (known as ‘pork-barrelling’) that they can.[66]

This also marks one of the greatest differences between those standing in Presidential and Congressional elections. If a President achieves little, they can claim they were blocked by other institutions, as Trump does with the ‘do nothing Democrats’.  If a congressional representative achieves little, they will be found personally responsible for their inactivity by their constituents and elected out of office. In this way, senators and representatives to the House have a much closer relationship to their electorate than the President, naturally, and hence the fate of congressional representatives is tied to partisan heavyweights like the President, Secretary of State, or the speaker of either legislative chamber for example.

Congress as It Stands

Currently, as can be seen from Figures 7, 10, 11 and 12 below in the appendix, the Republicans hold the upper house (with 52 seats) and the Democrats the lower (with 235) following the 2018 General Election. This spread of partisanship, as discussed in the introduction, has led to a considerable check on the power of the Republican party, as the 2018-2019 budget would not have been blocked, nor articles of impeachment drawn up, for example, had the Republicans retained the house. This spread is up for adaptation once again as all 435 seats of the lower house are to be fought for once again, and 35 Senate seats have come to the end of their six-year incumbencies. What could the result of this election mean for Congress and the separation of power within the US political system?

If the GOP loses only 4 of the 23 Senate seats it will contest, and the democrats continue to hold all 12 of theirs, the balance of power in Congress will shift to the Democrats. Paralleling this scenario, all 435 seats of the House of Representatives will be fought for, and thus there are five possible outcomes as far as the congressional balance of power is concerned (in no particular order):

(1) The Democrats strengthen their majority in the house alone, and the Republicans hold their majority in senate, in which case nothing really changes.

(2) The Republicans hold the senate and the Democrat majority in the house thins but remains a majority, in which case little changes despite a greater emphasis being placed on internal partisan politics to ensure legislation is voted for along party lines.

(3) The Democrats hold the house and gain the senate; in which case everything changes. If Biden wins then government will be unified, able to pass legislation with ease, and if Trump wins, government would be bitterly divided. In fact, I would wager that this would become the stereotypical illustration of divided government, with every act of each agent checked and blocked by the other.

(4) The GOP hold the senate and gain the house, in which case government is equally unified and the balance of power in Congress rests with the Republicans alone. Government, in this case would either be divided, if Biden wins the Presidency, or wholly unified, if Trump wins.

(5) The most unlikely of outcomes is that the GOP lose the senate yet gain the house, and the Democrats vice-versa. Here congressional division would still prevail in which case very little changes once again; it would just so happen that the two major parties will swap their control over the prerogative powers of each chamber.

With this in mind, in the event that these elections will not be a Democrat landslide, as the pollsters predict, and government is divided, the party that holds Congress may indeed ascertain the power to become the gatekeepers of federal legislative activity. This is what is at stake with the November elections behind the glitz and glamour of the presidential elections.

 This chapter will seek to lay out and explain what to expect from the Congressional Elections this November. Much like the previous section, my attention will be set on discerning where will be pivotal for both parties to defend and go on the offensive. This is somewhat simpler than discussing the Presidential elections because these chambers are divided into seats, and subsequently a magnifying glass can be placed over interesting races in their specific localities and constituencies, as opposed to focussing on the country as a whole. In order to do this, I will first examine The House of Representatives, undergoing the most change since 2016. This section will focus on overall trends since 1984, upholding the same timeline as my analysis of presidential elections, before discussing the 2018 election, and then examining a handful of interesting contests expected at this election. Following this, I will focus on The Senate, analysing overall trends since 1984 also, before being able to take an in-depth look at a number of the 35 contests that will be fought for in November. Finally, a conclusion and some final thoughts will summarise my thoughts on the distribution of power in Congress come January 2021 and which party will hold legislative power.

The House of Representatives

The first of the two chambers of Congress that I shall examine will be the House of Representatives. The chief function of the House of Representatives (the house) is to propose legislation and legislative amendments. Because of this function, control of the house is somewhat vital if a party wishes to be the gatekeeper of the legislative agenda. In this section I shall be discussing the electoral history of the house since 1984 with an eye to discerning which seats will be interesting or important to watch come November 3rd 2020. I will begin by discussing turnout before turning attention to how the distribution of seats and the popular vote has morphed across the last 36 years of elections, how the results of the 2018 midterms adapted the balance of governmental power, and lastly, which races to observe come the election.

Turnout to Congress

As discussed above, the house is elected every even year, bisecting the presidential term. As only one-third of the senate is elected at even years, midterms often tend to focus on the dynamics of the house, holding the potential to completely traverse the congressional balance of power, as happened in 1994, 2006, 2010, and 2018 (Figure 7). How does turnout in these midterms differ to that of the presidential elections however? Do the same number of voters realise the somewhat equal importance of electing the executive as they do the legislative branch of federal government?

Simply put, if we look at Figures 8a, b and c, we can see that the answer to this is a resounding no. In Figure 8a US General Election Turnout since 1984 is laid out in a cross-tabulation. The most obvious observation that one can detect is that turnout rises and falls with every general election. In years with Presidential elections, turnout is considerably higher than in midterms. The highest general election turnout since 1984 was with the election of Obama in 2008, with a 62.2% turnout (VEP); sharply contrasting to its lowest point only six years later, in 2014 – a midterm. In fact, if we look at Figure 8b, a graphic representation of Figure 8a, we see that no single midterm since the 1984 elections, even at its highest point in 2018 (50% VEP), is greater than the lowest turnout of Presidential elections in the same period, which was in 1996 with the re-election of Bill Clinton to office (51.7% VEP). Overall, the mean turnout (VEP) for general elections in this period was 48.99%. With a standard deviation (σ) of 9.17961, turnout has thus not fallen below -1σ, where midterms take an area under a normal distribution curve of 45.34%, but it has risen just above +1σ, where elections with a Presidential race take an area of 38.12%. Indeed, the only midterm to return a turnout higher than the mean was in 2018.

If we take a more thorough look at the data, is confirmed is that election turnouts increase and decrease like clear jagged waves on a turbulent sea. If one takes the data in Figures 8a and 8b and calculates the average distance between elections with presidential races and midterms, the average (μ) comes to an astonishing 17% difference.[67] The highest increase between a midterm and a general election with a presidential race was from the 36.7% low of 2014 to the election of Trump in 2016 with 60.1% turnout (VEP), a 23.4% difference. The greatest decrease was only the cycle before, from 2012 to 2014, in which turnout (VEP) dropped to 36.7% from 58.6% – a fall of 21.9 percentage points.

All trends seem to indicate that turnout will rise come November third. There are three reasons explaining such a prediction. The first concerns the sheer number of early voters and mail-in ballots. Already, as I am writing this, even though the election is less than one week away, over 75 million voters have already cast their ballot, surpassing over half of the total turnout of the 2016 election.[68] If this trend continues, and with the usual surge expected on election day yet to come, all indications point to this being a historically large electoral turnout. Beyond this, returning to the data, secondly, as discussed above, if one is measuring the turnout from the last general election (2018), the trend is clear, it will increase further, as every presidential election does from the previous midterm. Thirdly, and lastly, if we look at Figures 8b and 8c, turnout is generally increasing with a small positive linear association across time. Although predictions on the final turnout are varied and numerous, what we do know for certain is that it will rise – it is just the extent of the rise that remains as yet unknown.

What I would like to suggest is that we should not forget the average turnout for federal elections since 1984 sits at 48.99%, with a σ of 9.1796, and that +1σ from the mean has only occurred 5 times in this period. With this in mind, although the higher turnout the better as far as democratic legitimacy is concerned, the probability of this increasing above +2σ (67.349%) returns a p-value of just 0.0228, and +3σ (76.539%) at p = 0.0013. For English speakers, this means that the probability of seeing a turnout (VEP) of 67.3% is roughly 2.28 times in every hundred, and 76.5% is roughly 0.13 times in every hundred. Perhaps we should be sceptical about the massive increases some project to see, or perhaps this will be an election that defies statistical probability drawn from past data? Nonetheless, what can we discern from partisan seat distribution within the house?

Trends and Seats in The House

Assessing how the distribution of seats has changed since 1984 is interesting, but perhaps even more fascinating is that given that there have been 18 general elections in this period, the balance of power has only shifted in the house 4 times: 1994, 2006, 2010 and 2018 (Figures 7a and 7c). What unifies all of these is, firstly, that they are midterms, and secondly, that they are midterms after a president is elected to office for their first term (Bush Jr aside) where control of the house tips to the opposition party to the incumbent president.[69] If we look at Figures 7a and 7c more closely, we can see that the house holds a considerable majority of seats for one of the two parties consistently for at least two election cycles. From 1984 to 1994, the Democrats controlled the house with a majority, then the Republicans until 2006, the Democrats for a short two terms, and then the Republicans held the lower house until 2018 when the Democrat total number of seats jumped from 194 to 235.

The highest number of seats won by a single party in this period was in 1990, the midterm election of what was to become the one-term administration of president Bush Sr., where the Democrats won a total of 267 seats. The lowest number, naturally as a result of the two-party system, being the same year, some 167 seats won by the Republicans. The tightest difference came with the 2000 election, where the Republicans won a majority of just 9 seats.

Across the period, the number of seats won by each party does fluctuate with no obviously discernible trend concerning increases or decreases in conjunction with an incoming president. For example, in 1992 with the election of Clinton to office, although the democrats controlled the house their share of the seats decreased by 9. This compares to a similar scenario in 2008, where the Democrats controlled the house and a new Democrat president was elected into office, Obama, but here the Democrats gained a 24 seat increase. The same can be said for Republican Presidents, where the election of George W. Bush came with a decrease of 2 seats, but his re-election in 2004 came with an increase of 3. In this sense, as far as the number of seats any particular party gains, there is a lack of a clear trend as to the magnitude of seat changes with each presidential cycle. The only trends, as far as seats are concerned, relates to the first midterm after a new president is inaugurated, in which it tends to swing to the opposition party of the incumbent president, and secondly, that the balance of power in the house is yet to swap within one congressional election cycle of already doing so. I thus suggest that as it is not a midterm, and as the Democrats won back the house in in the last election, the Democrats will hold the lower house of congress once again. What about the popular vote and partisan vote share?

The Popular Vote and The House

As far as vote share for the two parties is concerned in the house, as can be seen in Figure 7b and 7d, since 1992 there is a tendency for the two-party system to be absolutely apparent, with the vote share of the two major parties almost totally mirroring one another. The highest share won since 1984 was in 1986 by the Democrats, where they received 54.12%, and the lowest by the Republicans in 2008 with 42.38%. As it stands, in 2018 the Democrats returned 52.9% of the vote, and the Republicans 44.26%. But what trends are there? How does this correlate to seats?

Apart from in 1996, where the difference in vote share between the two parties was at a low of just 0.3%, and 2012, returning a difference of only 1.31%, the party that has won the greater share of the popular vote has won the majority of seats in the chamber. Therefore, we should expect that if the popular vote in the upcoming congressional election returns a larger Democrat vote share, the Democrats should hold the chamber, following the pattern as discussed above in relation to seats. If we were to take an average of the two party’s vote shares since 1984, we would see that statistically the Democrats return a higher average than the Republicans, with 49.05% and 47.04% respectively. Conversely, we can see with a regression analysis that the Democrats are decreasing their share of the popular vote over time.  

If one looks at the graphic in Figure 7d, what one sees is almost a mirror after 1996, with every return falling between 41-55% of the popular vote. This presents the illusion of stalemate across the period, where it would seem that neither party gains or loses on the whole. Regression analysis reveals a different story. The Republican vote share since 1984 in elections to the house reveals a regression coefficient of ŷ = 0.05533x-63.6808, a correlation coefficient (r) of 0.21248, and a coefficient of determination (r2) of 0.04515. This implies a small positive linear association between time and Republican vote share – meaning that if we take the whole time period into account, the Republicans are statistically increasing in electoral popularity.

The same cannot be said of the Democrats, even though they have held the greater average share of the popular vote since 1984. By applying the same regression analysis to the Democrats, one can detect a regression coefficient of ŷ = -0.10682x+262.79975, where r = -0.34307, and r2 = 0.11770. This discloses that the reverse of the Republicans is true for the Democrats. The Democrat vote share returns a small negative linear association – meaning that if we take the whole time period into account, the Democrats are statistically decreasing their share of the popular vote and at a greater rate than the Republicans are increasing.

 This conveys two interesting phenomena are occurring side by side. The first is that although the electoral data may appear to be symptomatic of a wholly stagnant two-party system, with some exploration one may grasp that it is not stagnant nor mirror-like at all, but a two-party system in which the major parties are expanding or contracting their vote share at different rates. Secondly, perhaps more interesting, is that these rates are in reverse to the ones we saw for Presidential elections. As discussed in the previous chapter, as far as Presidential elections were concerned, the Democrat’s vote share is increasing at a greater rate than the GOP’s is decreasing. Although their coefficients of determination are dissimilar, the general trend of partisan vote share for House elections is the reverse of that for Presidential elections.

Overall, if this trend continues in this manner, in the long run we should expect the Democrat’s vote share to decrease, and the Republican’s increase – turning directly into seats, as discussed above. Whether or not we will see this at this election is another question, as all indicators point to another Democrat controlled house come January 2021, but over the course of future elections perhaps this will become more apparent.

The 2018 House Elections

In this section, I intend to talk a little about some of the more interesting returns from the 2018 midterm election for the house. I do not intend to linger for very long here, as a number of results have already been discussed in the previous chapter concerning Presidential elections. Equally, I have already presented why the Democrat win of 2018 was important, viz-a-viz congressional balance of power. I feel it would be counter-productive to repeat such findings, no matter how interesting they might be. Therefore, here I will lay out some broad patterns and comparisons between state-wide outcomes of the 2018 election, sparing the reader details of 435 races. Once this section is completed, I will draw up a list of outstanding congressional districts for the upcoming vote on November 3rd. The following will draw primarily on Figures 9a and 9b.

The Democrats stole the house in 2018 with 52.9% of the popular vote, gaining 41 seats. The Republicans, conversely, lost 42 seats and just over 4% of their vote share from 2016. As a result of the 2018 general election the house was divided; the Democrats held 235 seats and the Republicans 199.  But where were these increases in popularity for the Democrats hailing from?

As far as the percentage share of the popular vote is concerned, if we compare Figure 9a to Figure 9b, we can see the GOP lost a portion of its vote share from 2016 as a response to the Trumpian phenomena no doubt. This hardly came as a shock. Nonetheless, there were seven exceptions to this current. Alaska, Indiana, Rhode Island, Hawaii, Massachusetts, Connecticut and Wyoming all saw increases in popularity for the GOP. In Alaska, Wyoming, or even Indiana, this was understandable because all three of these states voted in majority for Trump in 2016 (Figure 5). Rhode Island, Hawaii, Massachusetts and Connecticut however were surprising. Rhode Island has elected only Democrats for the house since 1995, and the last time it popularly voted to award a Republican it’s ECVs was with Reagan in 1984. This being said, the Republican vote was still low at just 34.78%. Connecticut and Hawaii similarly so, and in Massachusetts, where the GOP increased its vote share by almost 5%, this still only pushed their total share to 18.09%. Perhaps this is will be inconsequential, but what it does show is a small demographic of the electorate that were swayed to vote Republican between 2016 and 2018.

In my treatment of popular vote and vote share there are four states that I have yet to discuss which specifically returned really rather dramatic decreases for the GOP from 2016. The most significant decrease came in Washington state. Although Washington usually affords its ECVs to a Democrat as far as most Presidential elections are concerned, the majority of seats won in Congressional elections can often swing like a pendulum. However, in 2018, the Republican party suffered a major loss, losing one seat and making their total number 3, against the Democrat’s 7, mirrored in that their vote share dropped by a staggering 14.96%. This was the largest drop for the GOP, but it was not overly surprising considering that only 36.83% of the Washington electorate voted for Trump in 2016, it followed a similar trend of firmly held Democrat states from 2016.

The more surprising were Kentucky, Louisiana, and Arkansas. In Kentucky, a red state through and through, the GOP still held 5 out of the total 6 seats, but lost 11.12% of the popular vote. If I were a Republican standing for re-election, as Senate majority leader Mitch McConnell (R) is, this figure would concern me. In both 2016 Presidential and Congressional elections, and the 2018 Midterm, the GOP returned a healthy majority of votes, as expected. Nonetheless, a decrease this great in Republican heartland indicates that there is a continually growing demographic of ex-Republicans who are not having their political grievances met. The same can be said, astonishingly, for Louisiana. Returning the same number of seats and almost a twin distribution of votes as Kentucky, in Louisiana the Republican vote share dropped by over 10 percentage points, and in Arkansas they lost some 8.59%.

 These are not just areas in which the GOP vote dropped, these are all states that voted at almost 60% or higher for Trump in the Presidential election of 2016 (Figure 5) and at around 70% in the 2016 Congressional elections (Figure 9a). Despite all being Trump-voting states, what makes these states interesting is that they are all part of the ‘Bible Belt’, with a high number of Christian voters – supposedly one of the greatest demographics to support the GOP. According to the Pew Research Centre, 56% of Protestant and 52% of Catholic registered voters lent their support to Trump in 2016, but a whopping 77% of evangelicals, of whom the bible belt is most often associated, voted for him also.[70] Perhaps this shift we saw in 2018 indicates that these figures will decline further?

Aside from this, where were the greatest gains and losses for the two parties? In Figure 9b what we can see is that alongside their drop in vote share, the GOP did not gain a single seat. In fact, all in all, the Republicans lost 42 seats, all of which the Democrats picked up.[71] The greatest loss came in California, where the Republicans decreased their share of the state’s 53 seats from 14 to just 7. Other loses came in New Jersey, Pennsylvania, Virginia, New York, Florida and Illinois, to name but a few.

Ultimately, this loss laid the foundations for the Democrats to control certain legislative processes of Congress with control of the lower chamber. The most important of these, of course, were the two articles of impeachment against Trump that passed the house with all but two Democrats voting in favour of both articles.[72] Now I have established what the outcome of the last elections to the house returned, I will briefly list a handful of the 435 house seats that will be of interest come November 3rd, which may decide if the Democrats increase or decrease their majority.

Which House Races to Watch

In this section, I intend to examine just a small number of seats that I see being at the centre of whatever little media attention the race for the House of Representatives receives. In many cases, some congressional districts have made this list simply as the 2018 result returned close results with thin margins of difference – indicating districts that could be fought for and won by either party.  I will lay these congressional districts out in numerical order within states that will be organised by an alphabetical logic. All of the following descriptive statistics have been taken from the compiled report by the Office of The Clerk of The U.S. House of Representatives.[73]


District 26 – In 2018 Debbie Mucarsel-Powell (D) won her initial term as a Congressional representative, becoming the first Ecuadorian-born individual to be elected to Congress and the first female representative of the district. The district itself is at the southern-most tip of the Floridian peninsula, in which all three of Florida’s national parks reside. Although taking this seat from the Republicans, the question remains as to the extent Congresswoman Mucarsel-Powell has achieved enough in her term of office, sitting on house committees for the judiciary, and transport and infrastructure. In 2018, she won with the thinnest of margins, gaining just 50.87% of the vote share with the then Republican incumbent gaining 49.13% – just a difference of 4,119 votes in a district where over 235 thousand ballots were cast.

District 27 – Donna Shalala (D) took the 27th Floridian district from a Republican hold, beginning her first term in 2019. Shalala had been an academic in her earlier life and became the U.S. Secretary of Health and Human Services throughout the Clinton administration. Since then she has held a unique connection to the 27th District, being President of The University of Miami until 2015. In her tenure as a congresswoman, she has thus far been a member of the House Committee for Education and Labour, and the Committee for Rules. Although securing a majority in 2018 with 51.76% of the vote share, and the GOP returning a full six points behind, at 45.76%, the question remains, much like Mucarsel-Powell, has Shalala has done enough to secure her seat? After all, the seat distribution in Florida between the Republicans and Democrats is 14-13 respectively, and as such, if the Democrats wish to take Florida in the house, they will need to hold every seat they can.


District 8 – In 2016, on the whole the rather well populated 8th district of Michigan voted in favour of Donald Trump, much like other parts of the state. Equally, in 2016, the district returned the incumbent Republican candidate to represent the congressional district for a second term. Come 2018, like 34 other congressional districts up and down the country, the electorate chose to give a female Democrat the opportunity to Represent them in the house.[74] In this case, Elissa Slotkin (D) was elected with 50.61% of the popular vote, with the Republican candidate not far behind her on 46.78%. Flipping between parties is a regular occurrence for this district in what I have already discussed to be a swing state. For Michigan, as it stands its 14 congressional seats are divided into halves, with the Republicans losing their two-seat majority in 2018. Thus, if either party wishes to make gains in terms of seats in the state to gain a majority, District 8 is where their gaze will turn.

District 11 – Much like the above, Michigan’s 11th district will be a site for a potential GOP gain if the Republican candidate campaigns well.  Currently Haley Stevens (D) holds the seat just outside of Detroit, taking it from the Republicans who had held it from 2013-2019. Stevens has a healthy majority in this case, winning 51.84% and with the Republican candidate almost 7 points behind. Somehow I doubt that she will lose her seat, but if the GOP can persuade the electorate that Stevens has achieved little in her term to benefit the district, be it a lack of pork-barrelling or being part of monumental decisions that chime with the electorate, they may be in with a chance of at very least thinning out this majority.


District 1 – On the southern border with South Dakota, Minnesota’s 1st district will undoubtedly be a site for competition at this election cycle. In the previous election, neither party made gains nor losses on 2016, but some districts did swap hands. In 2016 the seat was held by a Democrat incumbent and was taken by Jim Hagedorn (R). Interestingly, it was not an overtly huge surprise that Hagedorn won the seat after it had been represented by Tim Walz (D) since 2006. Indeed, since 2012, Walz was slowly losing a small share of the popular vote with every election, coming to its slimmest margin in 2016 with just a 0.8% difference in vote share. This thin majority did not change as a result of 2018, but the parties did indeed swap roles. Hagedorn won with 50.13% of the popular vote, and Dan Feehan competed as the Democrat nominee, winning 49.68%, thinning this margin to less than half a percentage point. With the increase in voters Minnesota returned for the Democrats in 2018, the Democrats will be eager to re-secure this seat.

District 2 – Angie Craig (D) took the constituency from the GOP, who had held it since 2001. Simply put, this too was a tight competition and the Democrats will be anxious to not let it slip, as doing so would allow them to retain their majority of seats for the state. Craig returned 52.66% of the vote share, beating her Republican competitor by 5.51%. Nonetheless, Craig should not simply expect to be returned automatically. As her win illustrates, the district that is just south of Minneapolis could U-turn if the Republican candidate can convince that 5 and a half percent of the electorate to once again place their faith in the GOP. If they were to do this, and the remainder of the state retain the same results as 2018, each party would hold 4 seats.

District 7 – The 7th congressional district in Minnesota covers almost the entirety of the western section of the state. Traditionally the district has historically voted Republican in Presidential races, but since 1991 Colin Peterson (D) has held the constituency. This goes to show two rather important points. The first is that the US electorate are not necessarily consistent in terms of partisan voting, and therefore support for any party in one mode of elections does not necessarily translate to support in another, requiring a deeper concern for demography if one really does wish to understand sub-state voting trends on an increasingly granular level.[75] Secondly, it shows the full extent as to how a strong representative-constituency link can lead to maximal pay-offs. This being said, Peterson was one of the two Democrats that did not vote in favour of the first article of Trump’s impeachment. One assumes that this was a political tactic in order to remain in favour with his majority Republican voting electorate in order to secure his re-election in November. However, if Minnesota turns against Trump even more so than it already did in 2016, this decision may bring Peterson’s almost 30 year career representing the district to its end.

New Jersey

District 2 – The 2nd district of New Jersey will be watched closely by the GOP. The incumbent of the seat is Jeff Van Drew, who won the seat as a Democrat in 2018, but has since defected to the Republican party. Like Colin Peterson in the Minnesota 7th, Van Drew did not vote for both articles of Impeachment in 2019. It was this decision that led him to cross the floor and turn red. This will be interesting as in 2018 he won with 52.9% of the vote on the Democrat ticket, with the Republicans securing 45.23%. Indeed, this 7-point difference is quite a gap to fill, but the question will really be the extent to which the electorate in New Jersey voted in 2018 for the representative or the party the representative is affiliated to. While we should expect Van-Drew’s popular vote share to drop, the question is whether or not it will be by feet or inches – whether he will lose those 7 points he cleared in 2018, or will hand them to the new Democrat nominee.


Districts 1, 10 and 16 – Although I spoke at length about Pennsylvania in the first chapter of this investigation, PA will be interesting to observe in these upcoming congressional elections. In early 2018, the state supreme court determined that Pennsylvania’s congressional districts were gerrymandered in order to benefit the Republican party, a practice that partisan figures in Pennsylvania have often, sadly, historically engaged in.[76] As a result of the 2018 elections, re-districting led to the Democrats gaining three seats by some majority. However, the combatting of gerrymandering did not lead to an outright win for the Democrats across the board. In three congressional districts (1st, 10th and 16th) the Democrat nominees won with no more than a 4% difference between themselves and the Republican candidate. What will be interesting to see is if re-districting in 2018 led to Democratic wins in response to the state Supreme Court’s decision that the Republicans had gerrymandered the state, or if re-districting has acted as a corrective and these constituencies are indeed populated by a majority of Democrat voters.

There are plenty of other seats that I could discuss. However, as the house contains 435 seats, plenty of which were close races in 2018, it would require a study in itself in order to go through every seat. The purpose of this section was but to give a flavour of the sorts of races we should expect to focus on, and to detail but a handful of those most interesting ones to watch come November 3rd.

The Senate

Of the one-hundred seats in the Senate, only 35 are to be fought for in November. This is more than the usual electoral cycle, with two special elections taking place in Arizona and Georgia. This means that both of Georgia’s senate seats are to be decided on, with 23 of the total seats being fought to be defended by Republican incumbents, and 12 by Democrats. Initially, what is at stake with this election is that the GOP could lose the few number of seats that afford them their majority in the upper chamber. If they were to do so, and the Democrats were to hold the house, then the legislative branch of government would turn wholly blue – i.e. the democrats would control the legislative process for the next two years, irrespective of who the President will be.

In this examination regarding the Senate, I will briefly focus on the wider historical trends of the chamber, before narrowing my focus to the patterns of this election cycle, and finally discussing some of the more fascinating races out of the 35 being fought for. This will conclude by claiming that the result is far from certain, either party could emerge from this election with the power of the upper chamber in their control. What is important are the number of seats the Democrats can contest to win. Currently, the distribution of seats sits in the Republican’s favour – holding 53. Nonetheless, the Democrats are not far behind with 45, and hold the support of 2 independent senators (Bernie Sanders – VT and Angus King – ME), bringing the division in the senate to 53-47. If the Republicans lose just 3 seats to the Democrats, the chamber, let alone Congress or government, will be divided. If this loss were to move to just 4, Congress will be won for the Democrats. This is what is at stake for the GOP. This election is theirs to lose in the Senate, and if they do so, they will hand the legislature to the Democrats. If Biden really is destined to win, as perhaps too many pundits are claiming so confirmedly, the Senate will be the only way to ensure government is not united by the Democrats – a feat that has not occurred since 2011.

Seat Distribution in The Senate

If one looks at Figures 11a and 11c, one will be able to observe how seats have been distributed between parties since 1984. In this time period, nine elections returned a Republican majority of seats, seven for the Democrats, and two stalemates. This implies that the Republicans have been slightly more favoured to win the house, but not by much. The Senate represents fully here the two-party system and the deadlock that US political institutions often find themselves in, with neither party gaining less than 40 seats. Thus, the senate’s typical condition is one of tight margins and close results. Every seat counts.

Since 1984, continuing with the same time period as my analysis of the House and Presidential elections, the greatest number of seats won was in 1992 with the Election of Bill Clinton to the oval office and in 2008 with the election Obama – returning 57 seats. The manner in which Senate elections function, whereby other parties outside of the two-party system are wholly locked out of the capability to compete, an increase for one is almost always inevitably a decrease for the other. Alongside this, it would be an oversight if I did not mention that the two independents sitting in The Senate have done so since 2006, in which their election led to the 49-49 stalemate between the two major parties. Yet, as explained above, these two Senators are aligned with the Democrats and are therefore decisive for Democrat control of the upper house.

In the time period examined, the balance of power has only directly changed hands five times, in 1986, 1994, 2002, 2008 and 2014 – a mixture of midterms and elections with a Presidential race. Much like the House of Representatives however, there has yet to be a period in which the power dynamics of the chamber have swung within a single election cycle. The Republicans have held the chamber for six years now, continuing their winning streak throughout a whole senate term. This means that the group of seats which tipped the Republican share into a majority in 2014 are up for election once again. So, what is the history of just these seats? The senate seats on this six-year-term cycle alone?

As far as seats are concerned, this cycle is particularly varied in how it may turn the Senate. In 1984, it led to a Republican majority of seats. In the next election after this, six years later, in 1990, it sustained a Democrat majority, then Republican twice, Democrat in 2008, and Republican at its previous scheduled election in 2014. In fact, what is interesting about this is that the past two shifts of the balance of power in the senate were caused by this cycle. As it stands between the Democrats and Republicans, 45-53, astonishingly this was the exact same distribution of seats, all be it reverse in terms of partisanship, out of the 2012 election, two years before this cycle tipped the balance to the GOP. Perhaps it could do the same in this election, but for the Democrats.

Although because there are a greater number of traditionally GOP held seats in this cycle, two-thirds of the elections since 1984 have led to or sustained a Republican majority, in 1990 and 2008, this led to or sustained Democrat majorities. If one compares seat distribution in the House over time (Figure 7c) and seat distribution in the Senate over time (Figure 11c) for the years of this senate-election cycle, aside from 1984, they match. In the last 5 election cycles for this group of senate seats, they have either swung to or held in majority that party either winning or holding the lower chamber.

With the analysis presented in the preceding section kept in mind, that the Democrats are more than likely to hold the lower chamber of Congress, this leads to two potential outcomes. The first is that this senate election swings the balance of power to the Democrats in the senate, and hence Congress as a whole. The second is that on the consideration that the Democrats do hold the house, and the senate remains in GOP hands, this 5-cycle long trend is broken. Either way, with the manner in which these seats have swung the chamber in the past, it is hardly clear from seat distribution history who this senate election will favour. What can vote share tell us?

Partisan Vote Share in Senate Election

On the whole, since 1984, the average percentage of the popular vote awarded to the Republicans was 46.21%, and 49.7% for the Democrats (Figure 11b). If we take a look at the vote share in graphic form (Figure 11d), much like the house elections, popularity can swing from election to election, as we can see from 2008 to 2014. The highest share of the popular vote recorded in the senate was in 2018, where the Democrats won 58.4%, and as such, the lowest share was the same year for the Republicans, who received just 38.8%, a trend continuing from 2014. As time has passed, the gap between the parties appears to be getting wider as it concertinas between the two after the 2000 election. By far, the 2000 election returned the closest margin of difference between the two parties, with only 0.08% points difference (Figure 11a) Since then we have seen large swings and even larger margins of difference, unlike at the end of the last century, in which popular vote share favoured the Democrats from 1984 to 1994, but both parties remained within 4-5 percentage points of one another.

What does a regression analysis reveal? If we begin with the Democrat vote share since 1984, we can ascertain that there is a small positive linear association, returning a regression coefficient of ŷ = 0.07622x-102.82427, and a correlation coefficient (r) of 0.2. However, because of the large margins of difference experienced since 2000, the data series held a coefficient of determination (r2) of 0.04032, indicating that the popular vote share for the Democrats may not actually hold any correlation to the passage of time, with data points quite a distance from the trendline. Nonetheless, the regression coefficient of this association does signal that the Democrat’s vote share is increasing over time in elections to the Senate, but equally that this association with the passage of time is minor or of statistical intrigue alone.

If we look at the Republican’s vote share since 1984, naturally as a result of the two-party system there is some sense of mirroring. Here, there has been a rather mild negative linear association between the passage of time and vote share, returning a regression coefficient of ŷ = -0.14751x+341.37 and an r of -0.46607. This shows that there is a mild negative correlation between the variables of time and vote share. What makes this data series interesting, in comparison to the Democrats’ above, is that it held an r2 of 0.217, indicating that the data points fell closer to the trend line, making it more a proportionate variation than the historic vote share for the Democrats. In simple terms, the Democrats are increasing their vote share, but this increase is minor and as there is a high variation, the probability of there being a thorough correlation is slimmer, whereas the Republican’s vote share is decreasing over time with a more precise and definite negative association. But this only refers to Senate elections generally. Do results differ for the election cycle that will be voted for in November?

The answer to this question is, intriguingly, yes – in fact it is the reverse. By taking into account only the years where these seats were up for election, we can see that it is the Democrats who are decreasing their vote share with time and where 25.2% of variance in vote share can be explained by the passage of time (ŷ = -0.14x+334.86, r = -0.502, and where r2 = 0.252). For the GOP in this cycle, nevertheless, there really was no meaningful correlation, only a tiny positive association but with a minute 0.8% of variance in vote share explained by the passage of time, thus this regression analysis can hardly be considered telling of any particular trend, be it either a full determinate trend of increase nor decrease (ŷ = 0.02x-9.08, r = 0.08959, and where r2 = 0.00803). In this way, what little trends of vote share and partisan seat distribution that we can detect from past elections does not tell us much about who will return a majority of seats, if at all – making this senate election more fascinating a spectacle than others.

Although I do not wish to make any predictions, what I will say is that this Senate election contains the largest grouping of Republican held seats. Therefore, as this election cycle has historically shown, it would only be logical for it to return a greater number of Republican voters than in 2018. As a result of this, I would suggest that the wide margin of difference that currently stands (almost 20%) will undoubtedly be reduced. The question is whether or not it reduces to the extent that the Democrats will be able to pick up the requisite seats to win a majority.

The Senate Elections – Which Races to Watch

In Figures 13 and 14 in the appendix, I have laid out the seats that are to be fought for come this election, the incumbent, the previous election result (where colour indicates the vote share for the Democrats and Republicans) and the name of the challenger who is assumed to hail from the diametrically opposite party of the party system, but not always of course. I have divided these seats into two large lists: (a) those that Democrats defend (Figure 13), and (b) those that Republicans defend (Figure 14). Here, I intend to discuss these races for Senate seats at a granular level, simply examining why a handful of these races may be of interest. I will engage in a closer examination of 14 seats in total, beginning with those that the Democrats hold, before moving to the seats defended by the GOP.

Democrat Held Seats to Watch

Alabama – In a 2017 special election, the incumbent senator Doug Jones (D) won by the skin of his teeth, retuning 49.97% and with the Republican candidate gaining 48.34%. On the surface, this was absolutely a win for the Democrats, but the reality is that it was a loss for the Republicans. Jones faced and defeated the Republican nominee Roy Moore, who would have of been thought to be a shoe-in for the Republican held seat that was once won by the previous US Attorney General for Donald Trump, Jeff Sessions. Nonetheless, the GOP lost the seat not because Jones won the electorate over on a platform of policy, but because Moore faced allegations of sexual assault during the campaign for the special election, making his electability questionable as all personality flaws and candidate characteristics in prospective Senators often do.[77] This will not be an issue at this election, with the former head coach of an in-state university football team having secured the Republican nomination, Tommy Tuberville, is expected to be vastly more popular than Jones, with even calls from the Democrats to consider it a lost cause.[78]

Michigan – Although the incumbent senator for Michigan, Gary Peters (D), won with a comfortable majority of over 10% of the vote share at 54.6%, what makes this seat interesting is who will be contesting Peters. John James (R), the GOP nominee for the seat, has contested a Michigan senate seat before in 2018. In this election, although losing by a 6.5% margin, James still collected 45.76% of the votes against veteran politician Debbie Stabenow (D), who has held the seat since winning the 2000 election.[79] James, as a result of this feat two years ago, is now widely known in the state and stands a chance of beating the lesser known incumbent.

New Jersey – Cory Booker (D) and Rik Mehta (R) will go head to head to find themselves in this Senate seat come January. The reason why this is on the list is because New Jersey has not elected a Republican Senator since 1972. Subsequently, probability and precedent tell us that Booker will win.

Oregon – Jeff Merkley (D) who won with 55.7% against the Republicans’ 36.9% in 2014 will be competing against Jo Rae Perkins (R) to remain the incumbent of the seat. Perkins, astoundingly, was found by The Washington Post to be a supporter of the ‘Q-Anon’ Conspiracy theory.[80] The conspiracy contends that Trump is leading an effort to indict high-ranking Democrats supposedly involved in child-sex trafficking and paedophilia, and is connected to other conspiracies such as the ‘Deep State’ theory. Although the times do seem bizarre, with the high vote share that Merkley won his second term with, in 2014, and with this perception of Perkins widely known, the Democrats should rationally hold the seat.

Virginia – Mark Warner (D) held Virginia in 2014 by a rather thin scope. In 2014, he won his second term representing Virginia in the Senate with 49.1% of the popular vote, just 0.8% more than his Republican opponent. This was a far cry from his 65.03% win in 2008.[81] The question this election will pose for Warner is whether or not his vote share will continue to decrease, and if this transpires to be the case, he will more than likely be removed from office because his margin of victory six years ago was as slim as a margin could be. This is a seat that the Democrats will be fighting for. If it swings red, they will only have themselves to blame.

Republican Held Seats to Watch

Alaska – The incumbent, Dan Sullivan (R), won the seat in 2014 by beating the previous mayor of Anchorage Mark Begich (D). Now Sullivan is running for a second term, but there is no certainty that he will win once again. In 2014, Sullivan won with just over 2% difference between himself and Begich and without an absolute majority, hardly a landslide. Now he not only seeks a second term following this close race, but his challenger is an independent named Al Gross (I), who is running with the nomination of both the Democrat and the Alaskan Libertarian parties. This might just tip Al Gross to victory and de-seat Sullivan.

Arkansas – Tom Cotton (R) is the incumbent of this senate seat in Arkansas. The reason why he is on this list is that one can be pretty certain he will win this Senate seat as it stands. This is because no Democrat is running against Cotton, with the initial nominee John Mahoney (D) dropping out of the race just several hours after the filing deadline. The only opposition to Cotton is the evangelical missionary Ricky Dale Harrington, Jr. (L), running on the Libertarian ticket. Without any formulation of predictive bias intended against Harrington, my assumption is that the Republicans will begin this senate election with one seat already won.

Colorado – Simply put, Rory Gardner (R) will face the previous two-term Gubernatorial victor John Hickenlooper (D) who was, by all indications, a successful and popular Governor of Colorado. Gardner won the election in 2014 with just over a 2% margin, but his allegiance to Trump over the course of the past four years may harm him in Colorado, a state that voted against Trump in 2016 and will be expected to do so again in November. I would not be surprised in the slightest if Hickenlooper takes the seat for the Democrats, with these two points in mind.

Georgia – Interestingly, both senate seats are up for election with this cycle. Johnny Isakson (R) resigned from his position as Republican senator of Georgia due to health reasons in 2019. Since then, Kelly Loeffler (R) has filled the seat. As a result of this, Isakson’s seat will be filled in a Special Election come November 3rd, alongside the Georgia Senate seat that is at the end of its six-year cycle. Although in this race David Purdue (R) should hold his majority won seat from 2014, Loeffler (R) will have to face a ‘Jungle Primary’ in which multiple candidates may run from the same party and a majority winner will be elected and with a run-off if there is no clear winner. Loeffler is not the only Republican candidate, facing opposition from Doug Collins (R), consequently dividing the Republican vote and bestowing Raphael Warnock (D) with an opportunity to win the seat.

North Carolina – Thom Tillis (R) won closely in 2014 with only a 1.5% difference between himself and the Democratic candidate. Since then, North Carolina voted for Trump and a Democrat Governor (Roy Cooper) in 2016, and reduced the GOP’s vote share for the house in 2018. Tillis’s opponent, Cal Cunningham (D), is showing to be ahead in the polls, emphasising how Tillis’s support for Trump has been ad-hoc, meaning that he may well indeed steal the seat in North Carolina.[82]

Montana – Currently, as a result of the 2014 election, Steve Daines (R) holds a healthy 17.7% majority over the Democrats. What will make this election interesting to observe is that Steve Bullock (D) the incumbent governor since 2013 will run against Daines (R). This will be fascinating to watch because it will reveal an awful lot about the Montana electorate, which are considered to be a traditional GOP leaning electoral group. If they reduce their support for Daines on the whole, perhaps they are not as traditionally Republican as many contend, but that there is hope for the Democrats to make further gains in future elections.

Mississippi – This race piques my curiosity because it will simply be a rematch of the 2018 special election that Cindy Hyde-Smith (R) won and Mike Espy (D) lost. This will be a great indicator as to whether or not one of the states in the Republican heartland has become more or less sympathetic to the Democrats since the drawing of the Articles of Impeachment against Trump in 2019. As the race will be fought between the same candidates, the variable of character differences between candidates of the same party is eradicated, meaning that any change we witness will more than likely come as a result of partisan attitudes. This can therefore be treated like a partisan survey of the MI electorate, which is a rare occurrence across the nation.

Louisiana – Much like the Special election in Georgia, although the incumbent Senator Bill Cassidy (R) won the 2014 election healthily, the election outcome will be decided by an open vote without pre-selected primary nominations that has led to a number of candidates from all parties. This may de-seat Cassidy, but I expect the Republicans will still hold the seat come January. Nonetheless, the electoral process makes it fascinating to observe.

In Conclusion

In conclusion, overall, this analysis seems to suggest that The House will be returned to the Democrats. Although across time their share of the popular vote is decreasing at a greater rate than the GOP’s is increasing, in contemporary times there has yet to be a scenario in which the balance of power in the house shifts parties in two consecutive elections. Although this could in fact happen, it is highly unlikely given the popular sentiment of the times. It appeared that the 2018 midterms were in response to Trump’s leadership, and as such was a referendum on the GOP. If this was the case, then The Republicans did not return a victory.

If the majority of voters that chose to vote Democrat in 2018 decide that the GOP candidates have earned their support once again in the second half of the Trump administration, then perhaps the Republicans would be in with a chance of re-claiming a number of the 42 seats they lost. In order to do this, I expect them to weaponize the failed impeachment attempt by the Democrat controlled house, contending that it made the US appear divided institutionally, justifying why a vote for the GOP is a return to stability.

Nonetheless, the Democrat candidates will focus on the failings of this administration and Republicans over the course of the last two years, from federal COVID-19 measures, to personal scandals, the Senate’s dealing with the impeachment, and so on. This being said, I contend that the results of the house election will tell us not necessarily how the US feels about Trump, but about how they view the two parties on the whole. For this, it will be a fascinating race to watch, even if the Democrats do retain power over the lower house.

As far as The Senate is concerned, although trends in Seat distribution tell us that the Democrats may in fact take the upper chamber of Congress, the sheer number of Republican seats up for re-election will make this an uphill battle for the Democrats. Control of the Senate will be vital for the GOP, as if the polls are correct and that Biden will win the Presidency and the Democrats the House, the only mechanism for the Republicans left to check the Democrats is The Senate.

As far as historical popular voting trends go, I found that the Democrats are on the whole gaining support across time and the GOP losing support. Although these were statistically weak correlations, what we can see by looking at this electoral cycle of Senate seats alone is that the Democrats are slowly decreasing their share of the popular vote. If this regression analysis extends into 2020, expect the GOP to win the senate.

Although I will not firmly make any predictions, I do have expectations. I expect that the Democrats will hold the lower house, and in the senate races, the Republicans will gain one seat from the Democrats (in Alabama), and the Democrats could gain three from the GOP (in Colorado, North Carolina, and potentially Montana). However, this will not be enough for the Democrats to secure control of the Senate alone and therefore it may come down to the election of the independent Al Gross in Alaska if the GOP are to be castrated of their majority in the Senate.

If one wishes to follow how the broad landscape of US politics is to change as a result of this election, the Congressional races will more than likely provide an increasingly granular and in-depth grasp of the changing American political environment for the next two years at least. This makes this election far, far more than just the race for president, but for the immediate future of the country as a whole, be it more Trumpian, even in the potential wake of Trump himself, or a return to a politics that subsisted pre-Trump. Whatever way it is sliced and diced, with the Congressional elections ahead – America will decide.


The Electoral Crossroads

The Trump phenomenon is either coming to an end or just beginning to establish itself as a norm of the American political landscape. Writers and journalists alike have critiqued Trump for his temperament, his policy initiatives, his choices of staff, his broadcasted prejudices, his aesthetic tastes, and so on.[83] Equally, it seems to have become a tradition that when Trump inevitably rids the White House of a staff member but only a few days, weeks, months or years after joining his team, they write a book of memoirs in order to critique Trump the man for his character flaws, and Trump the president for his policy decisions.[84] When all is said and done, it will not be the writers and journalists that historians of US politics will undoubtedly associate with deciding the outcome of the 2020 election, but the American citizenry itself.

The purpose of this investigation was to highlight and explore where in the United States would matter more than elsewhere in the making of his historic decision. This required that its structure be broken down into two chapters, detailing each one the democratic elections associated with the two wholly democratically legitimate branches of government: the presidential election for the executive, and the congressional elections for the legislature. My interest here was not in predicting who or which party would ultimately emerge victorious come January 2021, but to discern which localities of the electorate could make a fundamental difference to the outcome of these federal elections.

In the first chapter concerning the Presidential election between Trump and Biden, I steered clear entirely from giving any kind of prediction. Nonetheless, I did draw up several possible outcomes that may arise. These were as follows, in no particular order:

1) Biden Wins with the kind of support that the polls are suggesting. The first potential outcome is that Biden will win with such a large share of the popular vote, an absolute majority if the current polls are correct, that one should expect Biden to secure 400 ECVs, or perhaps even more.

2) Biden wins with a higher share of the popular vote than Trump, as historical trends have set a precedent for, but not necessarily in all of the key states. For example, if this election were to return the same result as 2016 in all but Iowa, Michigan, Pennsylvania and Wisconsin falling into the hands of the Democrats, this would produce an outcome of 302 – 236 with Biden winning, a similar spread to 2016 but with party roles reversed.

3) Trump wins by securing the same states as in 2016. As far as my own predictive horizon scanning is concerned, I do not see any more states willing to give ECVs to Trump. If this does occur than there has been a serious overlooking of some variable by pundits and political scientists alike that would lead to any demographic of US society awarding some new allegiance to Trump and his Republican party. In this scenario I would expect a similar division of ECVs to 2016 but with the Democrat’s vote share increasing still, perhaps even with 4 or 5 million between the candidates, but only in states that the Democrats already held, like California, New Jersey or Nevada – effectively leading to yet another election where the Democrats win the popular vote but not the Electoral College.

4) Neither wins. There is always the prospect with such unpredictable elections that neither candidate will win. There is of course a slim percentage of the modelling that returns a 269-269 outcome, but it is highly unlikely. This being said, if it does occur, then the 12th Amendment to the constitution would come into effect and the House of Representatives would undertake the process of electing the new President by voting in state groupings, and the Senate would elect the Vice President.

Alongside this, a number of key historical trends were found by analysis electoral statistics from 1984 onwards, chosen as it seemed to mirror the qualities of this election without being either too soon nor anachronistic. As far as turnout was concerned, although it was discussed that there really was not a relationship between turnout levels and partisan pay-off on the national level, turnout in both terms of VEP and VAP have been slowly rising over the course of the past 36 years. Alongside the fact discussed in Chapter II that turnout always increases in federal elections with a Presidential race, there is enough evidence, alongside the long queues already at the polling places as I type this, to indicate that turnout will indeed increase dramatically from 2018, or perhaps even from 2016. Nonetheless, following the analysis of turnout undertaken in the previous chapter, one should not quickly forget the odds of turnout (VEP) exceeding +2σ or +3σ, making a turnout higher than 67% statistically unlikely – but, on the other hand, exceptions do often prove the rule.

Aside from discussing the number of states carried by candidates from either one of the two major parties across the period, a brief quantative analysis of US electoral data revealed a handful of trends concerning partisan vote share. After conducting a handful of regression analyses, I unpacked the pattern in the data revealing that the Democrats have increased their share of the popular vote in Presidential elections since 1984 at a quicker rate of growth than the GOP’s is decreasing, despite the fact that both parties’ share of the popular vote at presidential elections has been falling since 2012. If this pattern extends into the November elections, we would see Biden win, however, nothing is certain in this election and we must remember that this is only a broad trend.

After this had been undertaken, I presented a brief explanation of not how but specifically where Trump beat Clinton in the 2016 presidential election. This included a discourse around some of the reasons as to why Trump won in the areas he did, broadly speaking, that were rural with higher rates of income disparity and self-identifying evangelicals. This set the scene for my discussion of which states, and indeed which counties, will be significant at this election. If we can say that there was a thesis to this chapter at all, it was that this Presidential election will be fought in: Wisconsin, Florida, Minnesota, Michigan, North Carolina, and, perhaps the most important of all, Pennsylvania.

All of this was prefaced with a discussion of populism as a phenomenal quality of Trumpian politics. Although some might feel that this would have been an attempt to smear the right in the US, such an assertion would be wholly untrue. The Trumpian camp is connected intricately with the phenomenon of contemporary populism to the point at which the norms they action in electioneering, such as project-Alamo, are an extension of this phenomenon of contempt for liberal democracy and liberal principles as a whole. If one really does wish to grasp how the Trumpian Republicans grasp these elections, one must begin with the norms they profess to holding and enact through policy; their normative praxis presents the analyst with a doorway into grasping how they would perceive the current constitutional elections process, the electoral college, the processes of the legislature, and so on.

Once the presidential elections were discussed, I turned my attention to the Congressional elections taking place also on November 3rd. I cannot stress enough how important these elections also are for the future of the US. Even if Trump were to win the election, which is still a possibility, if he did not have control of Congress, he could find himself against another Impeachment scandal or blocked federal budget, and this time he may not have one half of Congress, The Senate, in his partisan favour.

This chapter began by laying out how Congressional elections differ to presidential elections, before engaging in separate dialogues concerning The House of Representatives and The Senate. Simply put, statistically speaking, the House is more than likely to remain with a Democrat majority even though there is a small negative linear association between the passage of time and their vote share, decreasing at a greater rate than the GOP’s is increasing – a juxtaposition to the trends we saw in the first chapter. Indeed, the House has only shifted partisan control a handful of times in the past 36 years, almost always only at midterms after a new president has been inaugurated, to the incumbent president’s partisan opposition, and, perhaps most importantly, the balance of power is yet to turn to tip after already doing so at the previous election. For all of these reasons, statistical reasoning and precedent tease the probability of a Democrat majority in the lower chamber of Congress.

Succeeding this, a number of key congressional district races were examined on an increasingly granular level, unpacking which races will be of importance both in terms of legislative arithmetic, but also in grasping the electorate’s mindset at this juncture.

As far as the Senate is concerned, following on from this, I discussed how trends in seat distribution over the course of the period reveal how the Democrats may very well take the upper chamber of Congress from the GOP also. Nonetheless, the sheer number of Republican seats up for re-election will make this a difficult task for the Democrats, and precedent is against them, with this group of Senate seats aiding or even making a Republican majority at two-thirds of its elections.

Nonetheless, as far as historical popular voting trends may reveal a pattern or direction of popularity that can be extended into the future, I assessed that the Democrats are on the whole gaining support across time in Senate elections, and the Republicans losing it, following a short regression analysis. This being said, these were statistically weak correlations that reversed when the same methods were employed in order to examine the voting trends of this particular group of seats, illuminating that the Democrats were steadily decreasing support – casting a sense of doubt that the GOP will lose its majority in the Senate.

Come what may though, the Republican majority will be fought for, as a victory for the Democrats would be to more than likely unify congress and the federal legislative process in their favour, and for the GOP it may well be their only way of ensuring that a potentially Democrat president may have their federal powers checked. I expect this to be a tight race and the 14 seats that I have listed at the close of this chapter to be decisive in determining which grand party will control the flow, processes and discourse surrounding the federal legislation the upper chamber must pass for any proposal to become law. The Senate election will be what makes the outcome of this vote either one of jubilation for the Democrats, or one of contentedness for the GOP.

There were five potential outcomes of these Congressional elections discussed. They are as follows in no particular order:

1) The Democrats strengthen their majority in the house alone, and the Republicans hold their majority in senate, in which case nothing really changes.

(2) The Republicans hold the senate and the Democrat majority in the house thins but remains a majority, in which case little changes despite a greater emphasis being placed on internal partisan politics to ensure legislation is voted for along party lines.

(3) The Democrats hold the house and gain the senate; in which case everything changes. If Biden wins then government will be unified, able to pass legislation with ease, and if Trump wins, government would be bitterly divided. In fact, I would wager that this would become the stereotypical illustration of divided government, with every act of each agent checked and blocked by the other.

(4) The GOP hold the senate and gain the house, in which case government is equally unified and the balance of power in Congress rests with the Republicans alone. Government, in this case would either be divided, if Biden wins the Presidency, or wholly unified, if Trump wins.

(5) The most unlikely of outcomes is that the GOP lose the senate yet gain the house, and the Democrats vice-versa. Here congressional division would still prevail in which case very little changes once again; it would just so happen that the two major parties will swap their control over the prerogative powers of each chamber.

These results permit us an understanding of the electoral context behind this democratic event, allowing us to see where will be decisive in determining, or perhaps even just nudging, a certain outcome. I would like to bring this investigation to a close by reminding the reader that these elections are fundamental to the democratic basis of the American political system as a whole. As the distinguished thinker Robert Dahl once argued, elections are the only feasible solution to filling the most basic requirement of the most elementary democracy.[85] Without elections, it would be thoroughly complex, if not impossible, to select en-masse who the founding father and co-author of ‘The Federalist Papers’, John Jay, once penned as: “the most distinguished by their abilities and virtue, and in whom the people have just grounds for confidence”.[86]

There have always been discussions of ‘responsible government’, but with elections and the awarding of democratic legitimacy to a package of principles and ideas, there is such a thing as ‘responsible citizenry’, whom take responsibility for the leaders they pick and the political programmes they legitimate.[87] Jay speaks of them ‘whom the people have just grounds for confidence’, and we must never forget that a vote for any given candidate is always perceived as a reasonable declaration of confidence in that candidate, whether one is protest voting, tactical voting, and so on. As far as confidence and responsibility are concerned, I may not be able to say who will win the elections come November 3rd, but what I can declare is that the US electorate has found itself at an electoral crossroad at every level and branch of government. Which way it will decide to go next will be its own decision and nobody else’s. All we can hope for is that there can be sense made of it all with the crossroad behind us and a clear path ahead.

But, however, it is time to make the choice plainly. America may decide, but for now, it stands at electoral crossroads.


Download PDF to see Appendix.


I would like to take this time to thank all those with whom I have had conversations with on the topic of US electoral history over the past several years. Thanks must be given to: William Mortimer, Sam Furlong, Emma Catt, Ben MacKay, Samuel J. Ascot, Dr. Stephen McGlinchey, Dr. Andrew Wroe, Thomas Newman and Chiara Marino. Conversations with each of you have been an immense help, even if you did not realise it at the time, nor how long ago they were. Thank you also to each of my students, your questions led to this analysis.

A special thanks must be given to James Pearce-Molland, who at such a young age has taken to the political more so than any other individual I have ever had the pleasure of meeting. His editorial work and discussion on this piece turned out to be more vital than he probably grasps. Thanks buddy.

[1] John Jay (1987) “LXIV: A Further view of The Constitution of The Senate in Regard to the Power of Making Treaties”, in Isaac Kramnick (Ed.), The Federalist Papers, London: Penguin Books, pp. 375-380, p. 376.

[2] Robert Dahl (2002) On Democracy, New Haven, CT: Yale University Press, p. 93.

[3] Leah E. Gose and Theda Skocpol (2019) ‘Resist, Persist, and Transform: The Emergence and Impact of Grassroots Resistance Groups Opposing the Trump Presidency’, Mobilization: An International Journal, 24(3), pp. 293-317.

[4] For a short and useful list of differing conceptualisations of democracy, see: Michael Saward (2003) Democracy, Cambridge: Polity Press, pp. 14-16.

[5] I have placed ‘we’ in parenthesis as the notion of what exactly constitutes a conceptual ‘we’ appears as being almost indistinguishable from an ‘us’, and as such, co-constitutes a ‘them’ – including within the boundaries and conceptual territory of a ‘we’ that which rests outside of it. Nonetheless to expand on this point would be to breach the scope of the current investigation. This justifies the parenthesis.

[6] Mabel Berezin (2017) ‘On the Construction Sites of History: Where Did Donald Trump Come From?’, American Journal of Cultural Sociology, 5(1), pp. 322-337.

[7] Hillary Rodham Clinton (2017) What Happened, New York: Simon & Schuster, Inc., p. 418.

[8] This measurement of COVID-19 infections was taken from: The New York Times (2020) ‘Covid in the US.: Latest Map and Case Count’, The New York Times,, /us/coronavirus-us-cases.html, (accessed 4th October 2020).

[9] M. de Secondat Baron de Montesquieu (1766) The Spirit of Laws: Volume One, Fourth Edition, Book XI Ch, 4, London: J. Nourse and P. Valliant, p.220.

[10] “When legislative power is united with executive power in a single person or in a single body of the magistracy, there is no liberty, because one can fear that the same monarch or senate that makes tyrannical laws will execute them tyrannically. Nor is there liberty if the power of judging is not separate from legislative power and from executive power. If it were joined to legislative power, the power over the life and liberty of the citizens would be arbitrary, for the judge would be the legislator. If it were joined to executive power, the judge could have the force of the oppressor”. M. de Secondat Baron de Montesquieu (2002) The Spirit of The Laws, Book IX Ch.6, Cambridge: CUP, p. 157.

[11] James L. Sundquist (1988) ‘Needed: A Political Theory for the New Era of Coalition Government in the United States’, Political Science Quarterly, 103(4), pp. 613-635, p. 622.

[12] Richard S. Katz (1980) A Theory of Parties and Electoral Systems, Baltimore, MD: John Hopkins University Press, p. 1.

[13] Giovanni Sartori (1976) Parties and Party Systems, Cambridge: Cambridge University Press, p. 188.

[14] For more information on the Two-Party System, see: Rod Hague and Martin Harrop (2013) Comparative Government and Politics: An Introduction, Basingstoke: Palgrave Macmillan, pp. 184-185; William Roberts Clark, Matt Golder, and Sonia Nadenichek Golder (2013) Principles of Comparative Politics, Thousand Oaks, CA: CQ Press, pp. 614-615.

[15] Daniel H. Lowenstein (2006) “Legal Regulation and Protection of American Parties”, in Richard Katz and William Crotty (Eds.), Handbook of Party Politics, Oxford: Oxford University Press, pp. 456-470.

[16] David R. Mayhew (1991) Divided We Govern: Party Control, Law Making and Investigations, New Haven, CT: Yale University Press.

[17] David McKay (2002) “Divided Governance: Does it Matter?”, in David McKay, David Houghton and Andrew Wroe (Eds.), Controversies in American Politics and Society, Malden, MA: Blackwell Publishing, pp. 9-19, p. 15.

[18] Cas Mudde (2004) ‘The Populist Zeitgeist’, Government and Opposition, 39 (4), pp. 541-563, p.543

[19] Cas Mudde (2019) The Far Right Today, Cambridge: Polity Press, pp. 123-124.

[20] George Hawley (2019) The Alt-Right: What Everyone Needs to Know, Oxford: OUP, p. 172.

[21] For more information on Project Alamo, see: Sue Halpern (2017) ‘How we used Facebook to win’, The New York Review of Books,, [Accessed 17th October 2020].

[22] Quoted from: Jody Avirgan (2016) ‘A History Of Data In American Politics (Part 1): William Jennings Bryan To Barack Obama’, FiveThirtyEight,, [Accessed 17th October 2020].

[23] Jamie Bartlett (2018) The People vs Tech: How The Internet Is Killing Democracy (And How We Save It), London: Ebury Press, p. 75-76.

[24] Michael Barbaro (2016) ‘The Best (and Worst) Campaign Ads So Far’, The New York Times,, https: // [Accessed 18th October 2020].

[25] Yochai Benkler, Robert Faris, and Hal Roberts (2018) Network Propaganda: Manipulation, Disinformation, and Radicalization in American Politics, Oxford: Oxford University Press. For more information on the topic of digital misinformation in election campaigns, see: Niels Nagelhus Schia and Lars Gjesvik (2020) ‘Hacking democracy: managing influence campaigns and disinformation in the digital age’, Journal of Cyber Policy, DOI: 10.1080/23738871.2020.1820060, pp. 1-16.

[26] The Washington Post (2020) ‘Donald J. Trump for President: Cards’, The Washington Post, https://www.wash–campaign-2020/2020/08/ 04/62412547-092d-41a7-a48b-9b96fc3a532f_video.html [Accessed 18th October 2020].

[27] The BBC News Visual and Data Journalism Team (2020) ‘US Election 2020 Polls: Who is Ahead – Trump or Biden?’, BBC News,, [Accessed 18th October 2020].

[28] Dana Blanton (2020) ‘Fox News Poll: Biden Gains Ground Over Trump’, Fox News,, https://ww [Accessed 18th October 2020]. Other polls that substantiate this point: FiveThirtyEight (2020) ‘Who’s Ahead in the national polls?’, FiveThirtyEight,,, [Accessed 18th October 2020]; Ashley Kirk, Dominic Gilbert and Bruno Riddy (2020) ‘US Election Polls Tracker: Will Donals Trump or Joe Biden Win 2020 Presidency?’, The Telegraph,, 2020/10/18/us-election-polls-tracker-2020-who-win-donald-trump-joe-biden/ [Accessed 18th October 2020]; YouGov (2020) ‘Presidential Voting Intention 2020’, YouGov,, topics/politics/trackers/presidential-voting-intention-2020 [Accessed 18th October 2020].

[29] FiveThirtyEight (2016) ‘2016 Election Forecast: National Polls’, FiveThirtyEight,, [Accessed 18th October 2020]. 

[30] This process is set out in Article II, Section I and Amendment XII to the U.S. Constitution.

[31] David McKay (2013) American Politics and Society, Eight Edition, Malden, MA: John Wiley & Sons, Ltd., p. 55.

[32] Thomas Holbrook and Brianne Heidbreder (2010) ‘Does Measurement Matter? The Case of VAP and VEP in Models of Voter Turnout in the United States’, State Politics and Policy Quarterly, 10(2), pp. 157-179; Michael P. McDonald (2020) ‘Why should I care if turnout rates are calculated as a percentage of VAP or VEP?’, United States Elections Project,, [Accessed 18th October 2020].

[33] David Houghton (2002) “Voting and Non-Voting: America’s Flawed Democracy?”, in David McKay, David Houghton and Andrew Wroe (Eds.), Controversies in American Politics and Society, Malden, MA: Blackwell Publishing, pp. 31-43, pp. 32-33; David McKay (2013) American Politics and Society, Eight Edition, Malden, MA: John Wiley & Sons, Ltd., pp. 120-121.

[34] (From Figure 3) These states include: Alabama, Alaska, Georgia, Idaho, Kansas, Mississippi, Nebraska, North Dakota, Oklahoma, South Carolina, South Dakota, Texas, Utah, and Wyoming.

[35] (From Figure 3) These states include: California, Connecticut, Delaware, Hawaii, Illinois, Maine, Maryland, Massachusetts, New Jersey, New York, Oregon, Rhode Island, Vermont, and Washington. 

[36] (From Figure 3) These states include: Arizona, Arkansas, Kentucky, Louisiana, Missouri, Montana, Tennessee, and West Virginia.

[37] (From Figure 3) These states include: Colorado, Florida, Indiana, Iowa, Michigan, New Hampshire, New Mexico, North Carolina, Ohio, Pennsylvania, Virginia, and Wisconsin.

[38] Daron Shaw and John Petrocik (2020) The Turnout Myth: Voting Rates and Partisan Outcomes in American National Elections, Oxford: Oxford University Press, p. 11.

[39] For studies concerning the demographics of Trump’s electoral base, see: Pew Research Centre (2019) ‘An Examination of The 2016 Electorate Based on Validated Voters’, Pew Research Centre,, https :// [Accessed 19th October 2020]; Mark Setzler and Alixandra (2018). ‘Why Did Women Vote for Donald Trump?’, Political Science and Politics, 51(3), pp. 523-527; Rachel Bitecofer (2018) The Unprecedented 2016 Presidential Elections, eBook, Cham: Palgrave Macmillan and Springer International Publishing AG; Stephen L. Morgan and Jiwon Lee (2018) ‘Trump Voters and The White Working Class’, Sociological Science, 5(10), pp. 234-245.

[40] Philip Gorski (2019) ‘Why Evangelicals Voted for Trump: A Critical Cultural Sociology’, in: Jason L. Mast and Jeffrey C. Alexander (Eds.), Politics of Meaning/Meaning of Politics: Cultural Sociology of the 2016 U.S. Presidential Election, eBook, Cham: Palgrave Macmillan, pp. 165-184

[41] For Ohio Poverty Rates by county, see: Rich Exner (2020) ‘Every Ohio city and county ranked for poverty, child poverty: census estimates’,, [Accessed 19th October 2020].

[42] Lauren Kent (2015) ‘Where do the oldest Americans live?’, Pew Research Centre,, https:// sures%2C%20Florida%20%E2%80%93%20which,highest%20percentage%20in%20the%20nation. [Accessed 19th October 2020].

[43] Pew Research Centre (2019) ‘An Examination of The 2016 Electorate Based on Validated Voters’, Pew Research Centre,, 16-electorate-based-on-validated-voters/ [Accessed 19th October 2020]

[44] Jens Manuel Krogstad and Antonio Flores (2016) ‘Unlike other Latinos, about half of Cuban voters in Florida backed Trump’, Pew Research Centre,, unlike-other-latinos-about-half-of-cuban-voters-in-florida-backed-trump/ [Accessed 19th October 2020]. 

[45] The White House (September 1st 2020) ‘Remarks by President Trump During Survey of Property Damage: Kenosha, WI’, The White House,, [Accessed 19th October 2020].

[46] David McKay (2013) American Politics and Society, Eight Edition, Malden, MA: John Wiley & Sons, Ltd., p. 88.

[47] Tova Andrea Wang (2012) The Politics of Voter Suppression: Defending and Expanding Americans’ Right to Vote’, Ithaca, NY: Cornell University Press, p.38. For more information on ‘voter suppression’, see: Richard L. Hasen (2020) Electoral Meltdown: Dirty Tricks, Distrust, and The Threat To American Democracy, New Haven, CT: Yale University Press;Spencer Overton (2007) Stealing Democracy: The New Politics of Voter Suppression, New York: W.W Norton Inc.

[48] Amy Du Pont (2019) ‘Wisconsin judge orders removal of 234,000 voters from state registry’, Fox 6 News Milwaukee,, [Accessed 20th October 2020].  

[49] Matthew Rothschild (2019) ‘Elections Commission is Right to Hold off Voter Purge’, The Capital Times,, d-off-on-voter-purge/article_6c74c336-9fe3-59a4-9edb-a78580b5f5a1.html [Accessed 20th October 2020].

[50] Lackawanna County (2012) ‘2012 General Election’, Lackawanna County,, =6 [Accessed 20th October 2020].

[51] Devin Dwyer (2020) ‘’Naked Ballots’ In Pennsylvania Could Be Election Wild Card’, abc News,, 4058 [Accessed 20th October 2020].

[52] Richard Hasen (2020) ‘How ‘Naked Ballots’ in Pennsylvania Could Cost Joe Biden the Election’, Election Law Blog,, [Accessed 20th October 2020].

[53] Federal Election Commission (2013) Federal Elections 2012: Election Results for The U.S. President, the U.S. Senate, and the U.S. House of Representatives, Washington, DC: Federal Election Commission, p. 6.

[54] Federal Election Commission (2001) Federal Elections 2000: Election Results for The U.S. President, the U.S. Senate, and the U.S. House of Representatives, Washington, DC: Federal Election Commission, p.12.

[55] Edward-Isaac Dovere (2016) ‘How Clinton Lost Michigan – and Blew the Election’, Politico,, [Accessed 24th October 2020].

[56] Federal Election Commission (2009) Federal Elections 2008: Election Results for The U.S. President, the U.S. Senate, and the U.S. House of Representatives, Washington, DC: Federal Election Commission, p. 6.

[57] Federal Election Commission (2013) Federal Elections 2012: Election Results for The U.S. President, the U.S. Senate, and the U.S. House of Representatives, Washington, DC: Federal Election Commission, p. 6.

[58] United States Census Bureau (2019) ‘Quick Facts: North Carolina’, United States Census Bureau,, https://ww [Accessed 24th October 2020].

[59] Pew Research Centre (2019) ‘An Examination of The 2016 Electorate Based on Validated Voters’, Pew Research Centre,, 16-electorate-based-on-validated-voters/, p. 11 [Accessed 19th October 2020]

[60] CNN US Politics (2019) ‘Live Election Results: North Carolina – House 9 Special Election’, CNN,, [Accessed 24th October 2020].

[61] North Carolina State Board of Elections (2019) ‘Voter ID’, NCSBE,, voter-id [Accessed 24th October 2020]. 

[62] Dianne Gallagher, Angela Barajas, and Chandelis Duster (2020) ‘North Carolina Appeals Court Blocks Voter ID Law’, CNN,, html [Accessed 24th October 2020]. For discussion concerning Voter ID law and its relation to turnout see: Lauren R. Heller, Jocelyne Miller, and E. Frank Stephenson (2019) ‘Voter ID Laws and Voter Turnout’, Atlantic Economic Journal, 47(2), pp. 147-157.

[63] For a tracker of early voting at this election and turnout predictions by state, see: Michael McDonald (2020) ‘2020 General Election Early Vote Statistics’, United States Elections Project,, [Accessed 24th October 2020]. The report mentioned in the text is referenced from:John E Greve and Maanvi Singh (2020) ‘US 2020 election could have the highest rate of voter turnout since 1908’, The Guardian,, -highest-rate-voter-turnout [Accessed 24th October 2020].

[64] Bruce Hoffman (2020) ‘Right-Wing Extremists: A Looming Threat To The US Election’, Council on Foreign Relations,, [Accessed 23rd October 2020].

[65] This limitation is set out in the 22nd Amendment to the US Constitution, ratified in 1951.

[66] David Mayhew (1974) Congress: The Electoral Connection, New Haven, CT: Yale University Press.

[67] Equation for μ, see PDF.

[68] Michael McDonald (2020) ‘2020 General Election Early Vote Statistics’, United States Elections Project,, [Accessed 28th October 2020]; Reuters Staff (28th October 2020) ‘US early voting tops 70 million, continuing historic pace’, Reuters,, nuing-historic-pace-idUSKBN27D04L [Accessed 28th October 2020].

[69] Despite that the 2006 midterm was in the middle of Bush Jr’s second presidential term, the first was in 2002 in which we see an increase in electoral support for the Republicans from the 2000 election, more than likely as a result of the securitising rhetoric surrounding the ‘War on Terror’ and broadcasted ‘success’ of US intervention in the Middle (Figures 7d and 11d). Considering that wars do indeed tend to disrupt electoral patterns, introducing new standards and measures of critique on a governing party, the 2002 midterm should be considered and explainable anomaly to the pattern detailed above.

[70] Pew Research Centre (2019) ‘For Most Trump Supporters ‘Very Warm’ Feelings For Him Endured: A Detailed Look at the 2016 Electorate, Based on Voter Records’, Pew Research Centre, file:///C:/Users/kiera/Downloads/8-9-2018-Validated-voters-release-with-10-2-19-and-10-17-18-corrections.pdf [Accessed 28th October 2020], p. 13.

[71] This figure of 42 losses for the GOP includes North Carolina’s 9th District that was made to re-run its election in 2019. As the Republican’s held the seat beforehand, I have counted it as a loss. The GOP then went on to reclaim the seat in 2019.

[72] The two Democrats that did not vote for both articles of impeachment were Collin Peterson (MN 7th), and Jeff Van Drew (NJ 2nd) who crossed the floor and joined the Republican party in late 2019 to fight the 2020 election as the GOP nominee for the district.

[73] Cheryl L. Johnson (2019) Statistics of The Congressional Election from Official Sources for The Election of November 6th 2018, Washington, DC: Office of The Clerk – US House of Representatives.

[74] Drew Desilver (2016) ‘A Record Number of Women Will Be Serving in The New Congress’, Pew Research Centre,, s/ [Accessed 28th October 2020].

[75] For more information on this, see: David Darmofal and Ryan Strickler (2019) Demography, Politics, and Partisan Polarization in the United States, 1828-2016, Cham: Springer Nature.

[76] Anthony J. McGann, Charles Anthony Smith, Michael Latner and Alex Keena (2016) Gerrymandering in America: The House of Representatives, The Supreme Court, and The Future of Popular Sovereignty, Cambridge: Cambridge University Press, pp. 9-14.

[77] Alan I Abramowitz (1988) ‘Explaining Senate Election Outcomes’, The American Political Science Review, 82(2), pp. 385-403.

[78] Burgess Everett and James Arkin (2020) ‘Democrats Leave Doug Jones Hanging as Senate Map Takes Shape’, Politico,, 73 [Accessed 25th October 2020].

[79] Federal Election Commission (2019) Federal Elections 2018: Election Results the U.S. Senate and the U.S. House of Representatives, Washington, DC: Federal Election Commission, p. 23.

[80] Felicia Sonmez and Mike DeBonis (2020) ‘Believer in QAnon conspiracy theory wins Republican Senate nomination in Oregon’, The Washington Post,, /believer-in-qanon-conspiracy-theory-wins-republican-senate-nomination-in-oregon/2020/05/20/bf2d910a-9aaa -11ea-89fd-28fb313d1886_story.html [Accessed 25th October 2020].

[81] Federal Election Commission (2009) Federal Elections 2008: Election Results for U.S. President, the U.S. Senate and the U.S. House of Representatives, Washington, DC: Federal Election Commission, p. 76.

[82] FiveThirtyEight (2020) ‘North Carolina: Latest Polls’, FiveThirtyEight,, https://pr [Accessed 29th October 2020].

[83] Two of the more well-known critical evaluations of Trump the man and the Trumpian administration are: Michael Wolff (2018) Fire and Fury: Inside the Trump White House, London: Little, Brown Book Group; Michael Lewis (2018) The Fifth Risk: Undoing Democracy, London: Penguin Books.

[84] To give two examples: Anthony Scaramucci (2018) Trump: The Blue-Collar President, New York: Hachette; John Bolton (2020) The Room Where It Happened: A White House Memoir, New York: Simon & Schuster.

[85] Robert Dahl (2002) On Democracy, New Haven, CT: Yale University Press, p.93.

[86] John Jay (1987) “LXIV: A Further view of The Constitution of The Senate in Regard to the Power of Making Treaties”, in Isaac Kramnick (Ed.), The Federalist Papers, London: Penguin Books, pp. 375-380, p. 376.

[87] For a greater discussion of political responsibility, see: Hannah Arendt (2003) ‘Collective Responsibility’, in Jerome Kohn (Ed.), Responsibility and Judgement, New York: Schocken Books, pp. 147-158.

[88] In this Vermont race, the incumbent Peter Welch (D/R) won the nomination for the Democrats and, somehow by some odd twist of fate, the Republican nomination also, winning on write-in ballots. He accepted the nomination for both parties but stood as a Democrat, and as such, the Republican ticket received no votes.

Published by K.J.O'Meara

K.J.O'Meara (Ba, MLitt) is an independent British political theorist whose academic interests include: The English School of International Relations, Arendtian Political Thought, Political Hermeneutics, Biopolitics and Critical Film Theory.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

%d bloggers like this: