Conventional political science wisdom since the 1950s has seen elections in America, and in other well-established democracies, as driven by the struggle for support of the 'median voter'. Well-oiled political party machines, arrayed not too far from the midpoint along a left-right spectrum, are widely thought to earn their victories through a struggle over the middle ground. Parties that stray too far from the centre risk electoral annihilation, or a protracted period in the wilderness. This was a comforting idea, and one used to explain why older democracies have not descended into chaos, and why newer democracies in Africa, Asia and Latin America have proved much more troubled. Without a neat left-right ideological anchor to permit competition for the centre-ground, democratic politics runs the risk of becoming the plaything of populist demagogues, or deteriorating into social crises or authoritarian rule.

The trouble with this vision, as expounded by Anthony Downs in his 1957 Economic Theory of Democracy, is that it assumes most citizens cast their votes rather than abstain, and it assumes that most agree about what the central issues are even if they embrace radically opposed stances. If so, the 'median voter' can be identified as the voter with the same number of voters to his or her right or left on any political spectrum. The theory does not work so well if — as in America — 40% or so of eligible voters do not participate in elections, or cast votes only periodically to embrace populist demagogues or promises of radical change. Turnout may be higher than in America in other parts of the world, but in Europe it has declined over recent decades, notably France and Germany. Where general elections or referenda are decided by the periodic emergence of those who do not habitually vote, many of the old assumptions of the median voter theory go out of the window.  

As the early United States US presidential election results came in from Florida on November 9th, some commentators thought they spotted a high turnout, and attributed this to angry Latin Americans mobilized in opposition to Trump's racist policies. As the scale of Hillary Clinton's defeat became apparent, others blamed low turnout among Democrat voters. In fact, turnout has been relatively low in US presidential elections ever since the 1920s and in 2016 it was close to the norm of the 1970s-1990s (Figure 1). What Barack Obama did was temporarily boost the Democrat vote, particularly in 2008, bringing many disenfranchised voters into the contest (Figure 2). The Republican vote fell slightly as compared with 2012, and Trump lost the election on the popular vote. Although many traditional Republican voters refused to back Trump, he nevertheless secured victory in the rustbelt states like Ohio, Michigan, Pennsylvania, where white working class voters felt left behind by economic globalization and social dislocation (Trump regularly uses the term 'working class', unlike the Clintons who prefer 'middle class'). Based on the popular vote, the Republicans have not had a landslide victory since the 1980s.

Figure 1

Source: http://www.electproject.org/2016g

Figure 2

With two large parties of roughly equal size, the median voter theory suggests that political competition will pull both parties towards centrist policy stances. Parties of both left and right can rest assured of the votes of radicals who have nowhere else to go. It is the wavering 'swing voters' who determine the contest, and their movement back and forth between the two big parties is thought to yield a healthy periodic change from one government to another, and to ensure that the party of opposition sees its path to power by winning over the floating voter. Essentially the same theory underpins the justification for a presidential as opposed to parliamentary system, and a 'winner takes all' electoral law based on single member districts (with or without the U.S. electoral college arrangements). This, so the supporters of such systems argue, translates those periodic swings into strong mandates for the victorious party. Yet in contemporary America, as in many older democracies, the share of switching middle-ground voters is much lower than the proportion who perceive themselves estranged from the political process, and refrain from voting at most elections. In many contexts, victory goes to the politician who can draw the disenfranchised into the political arena, even if only temporarily and even if based on extravagant and ultimately unfulfilled promises, or else it goes to the politician who can push those periodic voters back out again, so lowering turnout for a rival party. In the 2016 US presidential election, some of the Obama Democrats may have switched to the Republican Party, but Trump's victory depended primarily on sustaining the largely white 2012 Romney vote while watching Hillary Clinton lose Obama's 2008 and 2012 additional turnout from minorities.  

Median voter theory also presupposes agreement on the central issues dividing an electorate. In the 1950s and 1960s, the left/right cleavage was central, and stances on that political spectrum could fairly accurately be mapped onto voters' socio-economic backgrounds. The theory disintegrates if Democrats who support Bernie Sanders, as well as Republicans who support Donald Trump, want to trash trade deals and impose tariffs so as to 'bring jobs back home', or if radical Republicans and Democrats are prepared to align themselves behind candidates who fiercely oppose the 'Washington establishment'. The issues that fundamentally animated the 'Leave' and 'Remain' campaigns in Britain's EU referendum likewise entailed a blurring of the left-right divide, with support for barriers to immigration figuring prominently, as was the case for those mobilized behind Trump's 'build a wall' campaign.

If American politics, and politics in the older democracies, are no longer anchored by the Downsian 'median voter', that suggests the need for a change in the way political scientists think about electoral competition. The 'politics of the lesser evil' that inspired so many to line up behind centrist candidates to defeat a common rival risks losing its strategic rationale. In such circumstances, prospects for electoral victory, as well as dangers of demagoguery, lie either in energizing the forgotten, the cast aside and the estranged, or subduing them if they are likely to back a major rival. That may be no good thing for democracy, not only because it encourages fleeting charismatic politicians who fail to deliver on their promises, but also because older and well-established politicians react by sneering at those ordinary folk who threaten them with defeat at the polls and prefer to insulate themselves from electoral accountability. Yet it also offers a welcome challenge for the mainstream parties, perhaps better than chasing after the 'median voter', to deliver substantive change for working people.