What the state of the US economy can tell us about Trump's populist appeal
Can bad economic governance explain the global rise in populism?
The Theory of Economic Voting
‘Economic Voting’ refers to tendency of voters to re-elect incumbents when the economy is performing well. A popular and well supported theory in the social sciences, studies of the phenomena have come to link a range of macroeconomic indicators to the performance of incumbents, most notably GDP (positive), unemployment (negative) and inflation (negative).
With perceptions of these indicators playing a commanding role in their salience, by no means is the economy a wholly predictive determinant of incumbent electoral success. Nonetheless, when economic fluctuations are dramatic, the economy can be a powerful weathervane to inform our understanding of the direction of an election’s otherwise invisible currents.
A Year of Populist Upsets
I began working on my Master’s thesis in November 2023 with the agonising awareness that my work would be immediately outdated by the deluge of electoral results that were to come in the preceding months. Being the ‘largest’ year in democratic history, 2024 has had a historic 1.5 billion people voting in some 50+ elections, nominating state and national leaders to govern roughly half of the world’s population. This meant that my masters thesis, which was on the relationship between the economic governance of non-populist incumbents and populist electoral success, would lack valuable data on the otherwise rare-event phenomena of successful populist upsets.
Of particular interest to me, was the record-breaking amount of populist candidates and parties that stood a viable chance of snatching power in 2024. From the AfD in Germany, to Eurosceptic Bloc in the EU to Trump in the United States, many well-known populist movements were positioned to make gains this year. More devastating to the relevance of my research was the unexpected results this year saw, from a minority BJP government in India to a plurality victory by Netherlands’ Gert Wilders to the unforeseen gains of Jean-Luc Mélenchon’s New Popular Front coalition in the French parliament. All of these results could have been used to better inform my modelling.
The silver-lining to the unfortunate timing of my research, however, is that the conclusions I drew can be used to guide intuitions of the relevance of macroeconomic variables in our upcoming elections. In particular, with the recent publication of my thesis by Monash University and the University of Warwick, I am now free to discuss how my findings relate to the upcoming November 2024 American presidential elections.
Economic Determinants of Populist Success
Inspired by the 2004 work of Collier & Hoeffler in studying the economic determinants of civil wars, my thesis presents a relatively novel approach to estimating political outcomes; extrapolating the logic of how economic factors may motivate the political outcome of a civil war to the analogous case of a populist emerging successful in an election. In particular, I estimate a logit regression model on a panel of political and economic indicators from 105 electoral democracies between the years 1980 and 2020, attempting to establish whether changes in these indicators coincidence with changes in the success of populist candidates.
To give a simplified example (which I explore in greater detail in the paper), consider the case of Jair Bolsonaro’s rise to power in Brazil’s 2018 election. Between the years 2003-2016, Brazil elected Workers’ Party (PT) candidates Lula Da Silva and Dilma Rousseff to four consecutive terms in the Presidency. Being consecrated with the title of the “most popular politician on Earth” by President Obama, Lula Da Silva commanded an unheard of 90% approval rating upon leaving office in 2010. Rousseff entered her second term in 2014 with similarly high approval ratings, but by the next election in 2018, the PT’s candidate was savaged in the polls and lost the presidency to a party that had just one seat out of 513 in Brazil’s lower house.
Figure 1 - Brazil GDP growth, Inflation and Unemployment (2000-2020)
So what happened between 2014 and 2018 that made such a seismic political upset possible? As with any social science event, the causes of Bolsonaro’s election are confounded by layers upon layers of complex and interdependent social phenomena. Nonetheless, by restricting the domain of our question, we can draw some useful insights. In particular, let’s ask the more constrained question of “Did the state of Brazil’s economy benefit Bolsonaro as a prospective populist candidate?”
Without presupposing whether the state of the economy ought to be subjectively good or bad to benefit a populist, we can then test whether there is an observed relationship between the state of the economy and incidents of successful populist electoral bids. From this relationship, we can infer the direction of the observed effect and determine whether this effect is causal.
Rounding out our discussion of Brazil’s 2018 election, studies general suggest that there was a profound causal relationship between Brazil’s catastrophic 2014-2017 recession and the appeal of populist ideology, personified in the increased appeal of Bolsonaro. While a full discussion of this event is given in my thesis, a cursory glance at the state of the Brazilian economy (figure 1) gives us compelling evidence to intuit some relationship between Bolsanaro’s success and the rise in unemployment, inflation and the decline in GDP that preceded the 2018 election.
Building off of this logic of examining the state of the macroeconomy 5-4 years before the successful election of a populist, I use the aforementioned dataset and logistic regression model to determine which factors routinely coincide with the successful election of a populist at the national level. A change in a variable coinciding with the change in another variable may not imply either a causal relationship, nor even a relationship at all. However, by including controls and ensuring we have a sufficiently large dataset, we can generally preclude coincidences, statistical noise, and confounded effects. To that end, I began my inquiry by collecting data on all of the populist leaders that were elected in the world’s ~105 electoral democracies between the years 1980-2020. This was done by consulting the political science literature, most notably by finding leaders who were classified as populist under political scientist Cas Mudde’s definition of the ideational populist. In the end, I settled on 39 leaders from 27 countries. Unsurprisingly the list consisted of a rogues' gallery of the world’s most polarising politicians. Donald Trump, Jair Bolsonaro, Narendra Modi, Benjamin Netanyahu, Viktor Orban all made the list.
Figure 2 - Estimates of economic determinants
Notes: Standard errors are in brackets. Significance levels are given by + p < 0.1, * p < 0.05, ** p < 0.01, *** p < 0.001.Without burying the lede any further, I found limited evidence that links macroeconomic conditions to populist electoral success. The exception to this was annual GDP growth, which, unlike inflation, unemployment, corruption, GINI or trade openness, is significantly linked to the odds of a populist being elected. In particular, a 1% increase in GDP reduces the odds of a populist being elected by 12.5%, within the next zero to five years.
This supports the general consensus within the political-economic literature that populists perform well during periods of economic crisis. The most powerful example of this is probably that of Adolf Hitler, who emerged as Chancellor of Germany by, among other things, capitalising on the economic instability caused by Germany’s post-War recovery and the Great Depression. Similar examples can be seen in the wave of Eastern European populist leaders that were elected in the early 90s, during the economic transition from a planned economy to a free-market one. And similar still is the tsunami of populists that emerged in Europe following the 2008 Global Financial Crisis and the Eurozone crisis.
Nonetheless, my results struggle to link other commonly cited variables, like trade openness, austerity or economic inequality to the incidence of populism. This does not mean that these factors are irrelevant as they have clearly played a role in some regional or national level explanations for particular experiences of populism. Instead, it appears that they are factors that are not universally relevant, one way or another.
The Economy and Donald Trump
With the IMF recently upgrading its US forecast, and declaring the American economy the fastest growing in the developed world, it may seem odd that the incumbent Democrats are struggling to edge out their populist challenger in the polls. But in the context of my own research, and recent trends, this outcome should hardly be surprising. As figure 3 implies, an otherwise impressive GDP growth rate of 2.8% only reduces the probability of a populist being elected (relative to if the USA was in a recession) by ~10%.
Figure 3 - Graph of relative probability of a populist electoral (p) victory and GDP growth rate.
A (relative) 10% reduction in the appeal of a populist candidate may seem like a decent amount in an election with razor thin margins, however, consider two facts about how the relationship between economics and voter behaviour (with populist challengers). Number 1 is that my model suggests that economic indicators can only explain ~13% of the variation in electoral outcomes when a populist is on the ticket. Thus, there is very little incumbents can do in the economic domain to win over new voters in such an election. Number 2 is that America is an outlier in my dataset, with Trump entering office in 2016 at the height of the post-GFC recovery, and leaving office in 2020 in the midst of recession. This does not mean that American populism is counter-cyclical, merely that a) macroeconomic indicators are not salient predictors of the economic experiences of average Americans, b) American voters have decoupled their judgements from ‘economic voting’ c) Americans are ill-informed about the reality of the Biden and Trump economies.
The Death of Economic Voting
Despite the economy being the leading issue on the minds of American voters, the 2024 US election is shaping up to be one shaped less by prudent economic calculation and more by imagination. The surprising superpower populist candidates have, is how easily they are able to buck historical trends by disengaging the objective faculties of voters in favour of dissociated narratives. In particular, the anti-establishment and anti-elitist ideologies that guide leftist and right-wing populist parties alike, are uniquely effective at dismantling otherwise successful non-populist governments from opposition.
As authors Funke, Schularick & Trebesch note in their seminal 2023 paper on the topic, despite being worse economic managers (GDP per capita tends to be 10% lower when compared to plausible non-populist counterfactuals), populists leaders are significantly more likely to be re-elected both in government and re-elected from opposition, and are significantly more likely to exit office in dramatic ways. The apparent longevity of populists, despite their poorer than average economic performance, should serve as a wakeup call to the DNC that simply being America’s technocratic ruling party is an inadequate strategy for gaining or maintaining power. To this end, I do believe that the results of my paper are fairly novel in that they underplay the role macroeconomics plays in predicting the election of populists. Economic voting may be alive and well when non-populist incumbents face a non-populist opposition, but when an election is contaminated by an ‘elites vs. outsiders’ outlook, allusions to your party’s recent performance only serves as a poisonous reminder of your party’s culpability in maintaining the poorly received status quo.
In a post-populist nation, perception matters more than reality. Consequently, good economic governance is not an effective re-election strategy for non-populist incumbents when facing a populist opposition. In any case, more research is likely required before we can make a full assessment on the electoral relevance of macroeconomics on the incidence of populism.
For more on this topic, consider reading my now published thesis “Can bad economic governance explain the global rise in populism?”. Additionally, if the intersection between politics, culture and economics interests you, consider subscribing to this Substack and sharing the post.