A number of thought provoking US election assessments and reflections have been written, including Alex Burns’s reflections and related analysis of the potential benefits of “red team” analysis and how that might have enhanced the Clinton campaign (link). I also took a strong interest in the US election and have been pondering what occurred and what might be learned from these events, though I suspect political scientists and others will continue to analyse the result for many years to come (in contrast to the snap judgements often published as opinion pieces). This post briefly outlines some possible factors and related lines of inquiry that appear relevant and worth exploring.
1/ Social scientists need to further examine the social constitution of expectations
It is now commonly agreed that the Clinton camp and most analysts (especially Left-leaning analysts) were overly optimistic. An important question is therefore why were they overconfident? Why did almost all analysts believe Trump had little or no chance of winning the election?
In many cases it could simply be motivated reasoning (link) or myside bias (link), though motivated reasoning doesn’t explain why many Republicans got it wrong. There are also reports that people in Trump’s campaign team doubted that he had a chance of winning. Similarly, Alex Burns’s analysis points to the importance of correctly interpreting signals of change, which may in part be addressed by adopting an enemy or adversary’s viewpoint (i.e. through “red team” exercises).
However I should also acknowledge that Trump was in deep trouble when his polling sank after the release of the ‘Access Hollywood’ tape and numerous women accused Trump of sexual misconduct and assault, and more Republicans distanced themselves from his candidacy.
Still, to me what’s interesting, and potentially more important, is how and why a consensus emerged that Clinton had the election ‘in the bag’ and how this was reinforced. This consensus perhaps contributed to campaign staff and analysts not taking key signals of change (like those pointed to by Burns) and risks (like those pointed to by Michael Moore) sufficiently seriously. Jens Beckert’s analysis of how shared “cognitive fields” can emerge and disseminate appears particularly relevant. Clearly opinion leaders, the mass media and social networks all play roles in this along with with what data is given more or less emphasis and how it is interpreted. Additionally, in the era of social media and filter bubbles such ‘cognitive field’ dynamics may occur more readily and be harder to escape or avoid.
In contrast to the emphasis on cognitive biases and flawed information processing (from more psychological studies), the US election result suggests a need for a broader focus on the ways that expectations get socially constituted. That is, moving away from individual-centred explanations and more fully examining expectations as social phenomena.
2/ Moral psychology may provide insights into why Democrats underestimated the Trump threat
A number of explanations have been proposed for why Clinton lost. Many explanations focus on Trump’s success in states such as Michigan, Wisconsin, and Pennsylvania (within the “rustbelt” region) and, related to this, stronger than expected support amongst white working class voters. Clinton lost this voting group by a large margin. Others contend that this explanation is mostly wrong and argue that Trump won “because Hillary Clinton was less attractive to the traditional Democratic base of urban, minorities, and more educated voters” (link). This argument contends that the main issue was that Clinton failed to inspire enough Obama voters to also turn out for her.
Nonetheless, it does appear to be true that many Democrats failed to anticipate key election results such as: 1) that Trump’s populist politics (and related anti-globalist economic agenda) would win him the votes of so many white working class voters; and 2) that the “Obama coalition” didn’t hold as well as hoped and Clinton won key demographics by lower margins (e.g. younger voters, non-white voters); and 3) perhaps most surprisingly women’s voting was similar to the 2012 in terms of the overall percent that voted Democrat and Republican. Moral psychology may help us to understand why many people on the Left seemed unable to correctly interpret the threat level and related trends (see link).
Principles of moral psychology such as that “morality binds and blinds” – which is proposed by Jonathan Haidt in The Righteous Mind – might help to explain why. This principle asserts that whilst morality is helpful for cohesion (e.g. where a group shares sacred values and/or sacred objects) it is “devastating for communities whose purpose is the pursuit of truth” (Haidt 2013). In other words Haidt argues that moral communities (e.g. political parties) and groups bind themselves together in ways that block open-minded thinking and hamper epistemic vigilance.
Haidt uses the metaphor of the matrix (from the movie of the same name), arguing that people live in shared “moral matrices” which are a form of consensual hallucination. He argues that in partisan debates people taking different sides literally live in different (moral) worlds and consequently they struggle – and typically fail – to understand the other side. This may also have been a cause of Clinton’s labelling many of her opponent’s supporters as “deplorables” – though, undoubtedly, some of the fringe elements of Trump’s base are deplorable.
Related cognitive mechanisms may have produced motivated ignorance (or motivated skepticism) regarding the risks inherent to Clinton’s liberal/Obama coalition strategy and more importantly regarding the developing move of white working class voters away from the Democrats. (NOTE: in the United States “liberal” is defined in opposition to “conservative”, where a liberal typically means an advocate of social and political change as per a progressive policy agenda, e.g through government intervention, as opposed to the “conservative” desire to uphold traditional values and ways of life. See this useful discussion)
Haidt makes the even stronger claim that “when a group of people make something sacred, the members of the cult lose the ability to think clearly about it… The true believers produce pious fantasies that don’t match reality, and at some point somebody comes along to knock the idol off its pedestal”. The US election may spark these change processes.
3/ The result highlights authoritarian threats to liberal democracy
A key issue in green political theory is whether ecological challenges justify more authoritarian States and associated illiberal approaches to social problems/policy. Many “survivalists” in the early 1970s were motivated by an intense sense of impending ecological crises and argued – often reluctantly – that a more authoritarian State is required to address these issues and avoid worse future problems (which may demand even more draconian methods).
Trump’s campaign often seemed to be following the playbook of survivalist movements (as well as the populist playbook), although to-date he’s been more successful than those movements.
His speeches often conveyed, and sought to create, a sense of impending doom which demanded (so the argument went) the more authoritarian approaches he advocated.
For people who are concerned about threats to liberal democracies the US election points to the need to better understand why people embrace more authoritarian leaders and what might be done to intervene in these processes. This lesson will be especially important if the Republican Party falls into line and backs some of Trump’s more extreme policies, many of which are inconsistent with much of the Republican Party’s current political philosophy. Similarly, if environmental challenges intensify in the future – as many scientists expect over the coming decades – these challenges may also herald major threats to liberal democracies. Important related questions include: Why was Trump’s demagoguery more effective than expected? How can moves towards authoritarian states be prevented/halted (including where these are an overreaction to environmental/social problems)? And, regarding a related line of inquiry examined in environmental politics field, can liberal democracies be ecologically sustainable?
4/ Prevailing cognitive models strongly shape imagined futures
Returning to the issue of expectation formation noted in the first theme/lesson, the result seems to be consistent with other sociological theories of expectations in particular the theory that actors’ expectations are “anchored in prevailing cognitive models, which function as instruments for the construction of imagined futures” (Beckert 2016, p.9). A general example is the ways that theories are a tool that actors use to construct their expectations, such as the influence of economic theories on actors’ mental models and their expectations of future developments in specific markets and the economy (e.g. consider references to the ‘law’ of supply and demand, etc).
With respect to the US election a number of ‘theories’ shaped expectations of the election result, such as the “blue wall” theory (see Nate Silver’s critique of this), related arguments regarding the political implications of changing demographics in America (e.g. given ethnic minorities tend to vote Democrat), along with beliefs about the importance of “ground game” advantage, the galvanising effects many felt Trump would have on voter turnout on the Left, etc, etc. Such ‘theories’ operated as cognitive devices but turned out to be less robust than expected.
An interesting reflective article entitled “I was wrong” by Republican strategist and writer Liam Donovan conveys this well. For example he writes: “I was wrong to buy into the Dem mythology of organizational superiority and data dominance. These things are important – even climactic – on the margins, but meaningless if the dogs won’t eat the dog food”. He also argues that “I was wrong to assume that the Obama coalition was transferable to another nominee. I was wrong to buy into demographics as destiny in the immediate term, however daunting the latent numbers”. He also notes that despite his blinkered perspective he saw signs that he was wrong but mostly ignored them. (Donovan’s other piece “The Blue Wall Crumbles” is also an interesting read).
Potential lessons here are diverse but include the need to be aware of influence of such cognitive models and to scrutinise underlying cognitive models used by actors.
5/ Imagined futures can problematically bias the evaluation of ambiguous evidence
One of the interesting things I noticed when reading US news during the weeks prior to the election is that events were often seen as confirming the predicted outcome. An example of this was early voting by ethnic minorities like Latinos which was widely seen as a sign of the predicted Latino surge that many expected would help Clinton win. This appeared to function as a kind of confirmation bias where ambiguous evidence is subjectively judged to confirm expectations.
Beyond these five themes, the election result appears to be yet another case example of where the wrong expectations are given credence resulting – in this case – in false confidence which impairs strategic action. As Nate Silver put it on The Daily Show:
“If the news media had acted like Trump had a 30% chance of winning [Nate Silver’s FiveThirtyEight gave Trump a 30% chance, which was much higher than other analysts] … then I think they would have acted very differently. Frankly, if the Clinton campaign had believed Trump had a 30% chance of winning they might have spent more time in Michigan and Wisconsin and not Arizona and Ohio and other states like that.”
Related suggested reading
Beckert, J. 2016, Imagined Futures: Fictional Expectations and Capitalist Dynamics, Harvard University Press.
Haidt, J. 2013, The Righteous Mind: Why Good People are Divided by Politics and Religion, Penguin Books.