Why the current peak in populism in the US and Europe ? Populism as a deviation in the Median Voter Theorem

The current surge of populism in Europe and the United States calls for further analysis using public choice tools. In this article, populism is modelled as a deviation from the normal state of the median voter theorem. This study adds to the public choice literature by proposing a model of populism which is suited, not only to left-wing populism, but also to other forms of populism prevalent in Europe and the United States today. It is argued that, due to changes in the assumptions underpinning the median voter theorem, the operation of the model can be modified, and as a result surges of populism occur. Those assumptions concern: the political spectrum; the distribution of ideological preferences; sociological, psychological and historical factors; political party competition; and extreme political preferences. It is shown that the current peak of populism in Europe and the United States can be explained through a simultaneous change in all of these aspects, leading to a “perfect storm” of populism.


Introduction
With the election of Donald Trump, the Brexit vote, the new populist government in Italy, and the gains in votes and seats of populist parties in several other European countries, we are clearly witnessing a surge in populism throughout Europe and the United States (Inglehart and Norris, 2016, Oliver and Rahn, 2016, Rooduijn, De Lange and Van der Brug, 2014) 1 .However, scholars are still divided as to what may have caused it and how it can be addressed (Inglehart and Norris, 2016).While there is a vast body of literature on populism in political science, the public choice focus on the United States and European countries.Latin America and other developing countries are not the main focus here, as it is argued that populism in those countries has different characteristics, which benefit from being modelled differently (Kaltwasser, 2012).literature on populism is scarce2 , and so far has focused only on left-wing populism and excessive government spending (see, for example, Acemoglu et al, 2013, andDornbusch andEdwards, 1991) neglecting to study the types of populism which we are currently experiencing in the United States and Europe, where government spending is no longer the crucial factor.
The article takes as its starting point the median voter theorem (Downs, 1957 3 ) which, it is argued, can provide us with the necessary tools to understand the multiple aspects of this phenomenon.The median voter theorem predicts that competition between political parties forces them to converge on the centre ground.This article argues that the rise of populism can be modelled as an alteration in this theorem, as voters are now increasingly rejecting parties, politicians and political positions that are at the centre of the political spectrum, and instead electing proponents of anti-consensual politics.Populist politicians emphasise their outsider status, and position themselves at the end-points, or even claim to be outside, the traditional political spectrum.This article therefore argues that this can be modelled as a situation where the median voter bliss point, which would be reached through consensual policies, is being rejected by both voters and populist politicians.4Downs (1957) provides us with a set of five assumptions that need to be met for the theorem to hold.In a nutshell, those assumptions concern: the political spectrum; the distribution of ideological preferences; sociological, psychological and historical factors; political party competition; and extreme political preferences.Framing populism as an alteration in the median voter theorem allows us to explore the causes of the current wave of populism, by analysing why several of those assumptions have changed simultaneously.
This article therefore argues that we are currently witnessing a "perfect storm" of populism, where different factors have converged to create a peak.While there have been references to peaks in populism being caused by a combination of factors which need to be present simultaneously (Oliver and Rahn, 2016), the literature has not yet, to the knowledge of this author, offered a comprehensive explanation of what those factors are, in relation to the current peak.
The use of the median voter theorem offers a framework in which to model and understand the parallel factors that are, together, feeding the current surge in populism.
Given the multi-faceted nature of the concept of populism, this article will not use one single definition, but will instead draw on different definitions offered in the literature.This will include the definition offered by Mudde (2007), of populism as a political party or politician that meets the three following characteristics: anti-establishment (opposing the "elites" and depicting the establishment as corrupt); authoritarian (emphasising personal power of a charismatic leader and/or direct democracy instead of representative democracy); nativist (against multiculturalism, and often displaying xenophobia).The article takes into account Taggart's (2004) definition of populism.According to him, populism is a combination of five elements: opposition to representative democracy; representing a heartland of "the people"; lacking core values; linked to a sense of crisis; unsustainable in the long-run.This article therefore aims to contribute to the public choice and politics literature on populism in the following ways.Firstly, by modelling populism as an alteration in the equilibrium state of the median voter theorem, it aims to propose a framework on which theoretical insights on populism can be included and systematically analysed.Secondly, it attempts to expand the public choice analysis of populism.To the knowledge of this author, the public choice literature so far has focused exclusively on left-wing populism, and its impact on macroeconomic policymaking, with a particular emphasis on excessive government spending, leading to unsustainable public debt and/or inflation (Dornbusch and Edwards, 1991). 5This article seeks to fill that gap, by using public choice in a way that is also suitable to right-wing and non-partisan populism.Thirdly, this study contributes towards the analysis of why there is a peak in populism in Europe and the United States in the 2010s, through the consideration of political, sociological, psychological, temporally dynamic and economic factors.
The remainder of this article is organised as follows.The next section presents the model offered in this article.The following sections address, in turn, each of the assumptions of the model and how they are relevant to ongoing populist developments.The last section concludes.

The Model: An alteration in the Median Voter Theorem
The median voter theorem states that, if a certain number of conditions are met, a democratic setting will result in policies which respect the preferences of the median voter.Models and their assumptions are, by definition, an over-simplification of reality, and the median voter theorem is no exception -but it provides an optimal starting point or benchmark, from which to analyse what 5 The concept of populism has also been used with yet another different meaning in the public choice literature.In Riker's seminal "Liberalism versus Populism" (1982), populism is understood as the view that democracy should be primarily a mechanism for ensuring that people's views are represented -in opposition to liberalism, where democracy should be primarily concerned with ensuring that people's best interests are represented.This understanding of the term "populism" is not the one being used in this article.
is different from the model in the real world, and why.Holcombe (1989) compares the median voter theorem to the assumption of perfect competition in Economics -although an unrealistic assumption, it has been crucial to that discipline, in that it provides a benchmark against which real-world situations can be assessed.
Similarly, the median voter theorem will be used in this article as the benchmark -a perfectly functioning democratic system exempt of populism -against which the real word situation, in which populism is almost always present, but sometimes to a much greater degree than others, will be analysed.Naturally, none of Downs' five assumptions is fully met at any time.However, what this article focuses on is whether the current wave of populism is linked to the fact that some of the assumptions are currently failing even more than usual. 6Downs (1957) provides us with a set of five assumptions that need to be met for the theorem to hold: 1. Political space is one-dimensional, and can be ordered from left to right; 2. Preferences are single-peaked and have a normal, bell-shaped, distribution; 3. Voters' preferences are not altered by historical, sociological or psychological factors; 4. Political parties can move within the spectrum (except beyond the nearest party); 5. Extremist voters may abstain if they perceive the two main parties as too similar.
Despite the vast literature that followed, it is argued that Downs' five assumptions, in their simplicity, remain the most useful analysis of the pillars upon which the theorem holds.This is reinforced by the fact that subsequent literature mainly focused on either proving or disproving the theorem, rather than on exploring the five above assumptions (Holcombe, 1989).This article will therefore analyse how the breakdown of each of these assumptions contributes to an alteration in the equilibrium state of the theorem.
In its normal equilibrium state, the theorem predicts that the preferences of the median voter will be adopted by policymakers.At first sight this may appear not to be directly related to populism, as populism can be placed anywhere in the political system.However, the wider interpretation of the political spectrum used in this article, whereby (following Converse, 1964) it depicts, not only the variations in policy preferences between left and right, but also the different options available to a policymaker, makes it possible to model populism as a deviation from the median voter's bliss point.That equilibrium therefore represents, not only the middle of a left-right political spectrum, but also balanced policymaking which, on average, meets the best interests of a representative citizen.Populism, by contrast, does not seek to offer balanced solutions that satisfy all citizens, but instead extreme and/or unusual solutions, aimed at satisfying only the interests of a "heartland that represents an idealised conception of the community they serve" (Taggart, 2004, p. 274).
6 The "failure" of an assumption is understood as any deviation from the normal or equilibrium state of the model, rather than as a criticism of the model.
To the knowledge of this author, the only existing study which connects the median voter theorem with populism is the study of Acemoglu, Egorov and Sonin (2013), but the authors offer a model focused on left-wing populism in Latin America, which, it is argued, cannot be applied to the current wave of populism in the United States and Europe.Using the median voter theorem in a radically different way from the one proposed in this article, the authors propose that populism can be modelled as a situation where voters willingly chose politicians to the left of the median voter bliss point.This is because they fear that politicians may be captured by business lobbies, and will therefore chose politicians who clearly signal that they are not captured, by offering leftwing policies.The authors also allow for the same phenomenon happening to the right of the median voter, where voters are primarily afraid of politicians being too left-wing (for example, being "closet communists").The authors therefore assume that the median voter theorem is functioning, but that populism will cause a bias to the left (or right).
However, it can be argued that their model, while very useful to explain the type of populism that happens in Latin America, is not helpful to analyse the type of populism taking place in Europe and the United States.In particular, current populism in Europe and the United States is mainly right-wing (Mudde, 2004), and cannot realistically be explained by voters fearing that politicians have been captured by lobbies on the left or are "closet communists".Ongoing concerns which are key to populism, such as opposition to immigration or a perception of having lost out to globalisation (Inglehart and Norris, 2016), cannot be explained by this model.Moreover, recent forms of populism have been disassociated from left or right-wing ideologies (for example, the election of Donald Trump and the Brexit vote cannot be seen as biases towards either the left or the right).Therefore, this article adopts a different perspective, by seeing populism, not as a bias in the median voter theorem, but as an alteration in its equilibrium state, due to changes in the assumptions that underpin the theorem.

Assumption 1:
The political spectrum is one-dimensional "1.The political parties in any society can be ordered from left to right in a manner agreed upon by all voters" (Downs, 1957, p.142)The assumption of there being a one-dimensional political spectrum, along which all political issues could be neatly placed, was always an over-simplification, and has arguably become even less relevant with time (Albright, 2010, Giddens, 1994).Despite the counter-argument that the left-right dichotomy remains relevant, as the values represented by left and right change over time to reflect new issues and concerns (Bobbio, 1996, Budge et al., 2001), it is impossible not to acknowledge that there is currently a multiplicity of policy dimensions in European and American politics, which cannot easily be placed along a single spectrum.Today, a political compass has therefore replaced the one-dimensional spectrum in the work of several scholars (see, for example, Inglehart andNorris, 2016, andBakker et al, 2015).
Is the current peak in populism partly caused by the fact that this assumption is breaking down more than usual?This article argues that this is the case, due to a combination of long-term and short-term factors.Over the long-term, we can identify a gradual dismantling of this assumption over the 20th century, for the following reasons.Firstly, an improvement in living standards made economic issues less relevant to voters, as opposed to social and cultural issues, which are less easily modelled in a left-right spectrum (Inglehart, 1990).Secondly, growing social mobility and narrowing social cleavages gradually reduced individuals' identification with an ideology (Franklin, Mackie and Valen, 2009).Thirdly, a convergence in political views made left less distinguishable from right (Albright, 2010).It is argued that this contributes to the long-term trend towards more populism, as voters are increasingly interested in political parties which do not fit within traditional ideologies, and therefore lose their identification with mainstream centre-left and centre-right parties (Franklin, Mackie and Valen, 2009).
It is also possible to identify short-term factors leading to a breakdown of this assumption nowadays.In particular, issues which do not fit easily into the left-right spectrum have become predominant in Europe and the US over the past decade.One of those is immigration -although anti-immigration rhetoric is often associated with the right, the anti-immigrant sentiment is usually captured, not by mainstream right-wing parties, but rather by populist single-issue parties, which are often categorised as right-wing but do not actually share any other right-wing views on other issues (Mudde, 2013).Moreover, as noted by Inglehart and Klingemann (1976), any issues which are not purely economic or religious fit with difficulty in the left-right axis.Immigration worries have become prominent in Europe due to the ongoing refugee crisis, whereby conflicts in the middleeast led to a sudden surge in numbers of refugees trying to enter Europe (Hansen, 2016).In the United States, immigration patterns and terrorist incidents have led to xenophobic feelings against Mexicans and Muslims (Inglehart and Norris, 2016).This has distorted the traditional concerns of the left-right axis.
Another short-term factor is Euroscepticism.Euroscepticism has emerged as a populist trend, linked to the perception that the European project was led by the elites, without proper consultation of the people (Taggart, 2004).Here also, it is possible to identify a combination of factors, which, over the past decade, has led to rising Euro-scepticism and questioning of the EU.
Those include economic difficulties, which historically have always been correlated with Euroscepticism, and in particular the 2008 Financial Crisis (Serricchio, Tsakatika, and Quaglia, 2013).
They also include the exponential pace of EU integration following the adoption of the 1992 Maastricht Treaty, and associated concerns about a loss of national identity (Eichenberg and Dalton, 2007).Euroscepticism can be categorised as an issue which does not fit into the left-right spectrum, and is instead better suited to the authoritarian-libertarian spectrum (Marks and Steenerberg, 2002).Bakker et al (2015) identify three policy dimensions in Europe: the traditional left-right spectrum; the libertarian-authoritarian spectrum; and the EU integration spectrum.The issues of immigration and Euroscepticism can be said to fit, respectively, into the second and third spectrums.A long-term trend in the break-down of the left-right spectrum is therefore currently coinciding with a short-term rise in the importance of political issues not directly related to the leftright spectrum.This is leading to a more than usual breakdown in the assumption of onedimensional politics, contributing to a more than usual deviation in the equilibrium of the median voter theorem.As people lose their identification with mainstream centre-left and centre-right parties, they are more likely to look for alternatives, and therefore more likely to become interested in populist movements.

Assumption 2: Preferences are single-peaked and normally distributed
"2.Each voter's preferences are single-peaked at some point on the scale and slope monotonically downward on either side of the peak (unless it lies at one extreme of the scale)."(Downs, 1957, p.142)According to this assumption, for the Median Voter Theorem to hold, voters' preferences need to follow a normal, bell-shaped distribution, whereby the majority of voters is located in the middle, and the number of voters becomes smaller the further away one moves from the middle 7 .This assumption is reasonable for a traditional left-right spectrum, when we assume that voters prefer a more left-wing or right-wing set of policies depending on their level of income (as naturally more redistributive policies benefit directly the poor, while being directly unfavourable to the rich, who will need to pay more taxes).This leaves the middle classes as the majority in the middle.
Once we introduce additional spectra, as discussed in the previous section, this assumption is still reasonable.In a spectrum going from liberalism to authoritarianism, it can still be expected that a majority will have a balanced view, away from the extremes.The same can be assumed of a spectrum going from complete EU integration to opposing the EU altogether.However, the assumption breaks down when voters are willingly asking for radical or extreme policies in very large numbers (demand side) or politicians are willingly offering extreme policies instead of balanced policies (supply side).This will happen when there is a polarisation of society, where groups of people have very different views from other groups.Inglehart and Norris (2016) show that society in the United States and Europe is becoming increasingly polarised, between a group of liberal-minded voters who will usually be more well educated and younger, and a group of conservative-minded voters who will tend to be less educated, older and often male.The authors show that this demographic split is relevant both in the case of the Trump election, and in the case of Brexit.In such a polarised society, as predicted by Downs, the normal bell-shaped distribution will be replaced by two peaks, leaving a gap at the middle.In such conditions, a politician with balanced views will be less successful in attracting votes than a politician with more extreme, populist, views. 7The Median Voter Theorem can be understood, both as a representation of one individual's preferences, and as a representation of a group's preferences.I am focusing mainly on the latter when analysing Assumption 2, as that is the most relevant aspect for my analysis.

Assumption 3: Voters' preferences are not altered by historical, sociological or psychological factors
"The distribution of voters along the scale is variable from society to society but fixed in any one society.[Reference to footnote 16:] Actually, this distribution may vary in any one society even in the short-run, but I assume it to be fixed in order to avoid discussing the complex of historical, sociological, psychological and other factors which cause it to change" (Downs, 1957, p. 142) Downs is here assuming that voters are perfectly rational, by ignoring any historical, sociological and psychological factors that may disturb such rational behaviour.Downs therefore did not argue that such factors were irrelevant to voters' decisions, but only that they were beyond the scope of his analysis.In the absence of historical, sociological and emotional factors, voters' decisions would be purely rational, and voting behaviour would maximise each voter's utility.We could then expect voters to select politicians whose views correspond to theirs, and/or whose policies are beneficial to their selfish interests.Such a voting behaviour will, if the other assumptions are also met, ensure that the theorem holds.
Once we make those factors endogenous to the analysis, voting decisions are no longer purely rational.Voters are no longer only maximising their utility, and are no longer voting for the party closest to them in political space.Instead, their behaviour may be guided by emotions, influenced by societal pressures or simply hampered by incomplete information.Again, this assumption is never fully met in the real world, but this article argues that it is being met less than usual at the moment.
There is a wide literature on the sociological, historical and psychological causes of populism, and in particular its association with crises (Taggart, 2004, Laclau, 2005) but there is a lack of research on why these causes are particularly acute now.This section builds on that literature, to examine why the current "perfect storm" of populism is due to a greater than usual variability in voters' preferences.In particular, I will focus on sociological, historical and psychological reasons which lead voters to be currently particularly attracted by populism.

Sociological and Psychological Factors
Inglehart and Norris (2016) argue that a crucial reason for populism is dissatisfaction on the side of the voters, for both economic and cultural reasons.The economic reason is a dissatisfaction emerging from changing economic trends with a significant impact on society, including globalisation, job insecurity and inequality.The cultural reason is a "backlash" against the rise in progressive values since the 1970s, in relation to multiculturalism, gender equality and human rights (Inglehart, 1990).The authors find that there is more conclusive evidence to support the cultural thesis, but also admit that the two theses may be too interrelated to be entirely separated, indicating that both are valid.Elchardus and Spruyt (2016) examine the link between support for populism and psychological factors, including declinism.Declinism is defined as the feeling that society is getting worse, due to globalisation, multiculturalism, as well as changes in regulations, the environment and moral principles.This concept is clearly very close to economic and cultural dissatisfaction identified by Inglehart and Norris.The authors find a strong link between declinism and populist voting, and conclude that a perception that society is getting worse, is the main psychological factor that can be associated with populist voting.
Crucially for this article, Inglehart and Norris (2016) find that economic and cultural dissatisfaction have been rising gradually since the 1970s, and are now at an all-time high, particularly among the sections of the population that have predominantly voted for populist candidates in recent American and European elections (male, white and older).This trend supports this article's thesis that voters' tendency to act in an emotional rather than rational manner has reached a peak.
It is argued that this long-term trend is now coinciding with a short-term peak, following the 2008 financial crisis.Crises, in particular of an economic nature, are often at the source of populism (Taggart, 2004).Taggart further argues that what is relevant for populism is not crisis itself but instead the perception of crisis.Although the Great Recession peaked in 2008, the sense of insecurity and agitation that it brought about has arguably not yet waned, particularly as many of the jobs lost during the recession were replaced with less permanent and secure positions, and the so-called Euro-crisis prolonged the recession throughout Europe.Inglehart and Norris (2016) are puzzled by the fact that economic dissatisfaction is higher, not among the lower classes, but among the middle-classes.It is suggested that this phenomenon can be explained by an economic trend: in both Europe and America, growth has been relatively higher for the rich and the poor than it has been for the middle classes -this is known as the "Elephant Curve" (Milanovic, 2013).Elchardus and Spruyt (2016) indeed emphasise the link between "feelings of relative deprivation" and attraction to populism, which they measure through surveys asking respondents to indicate how they perceive their economic situation in relation to that of others, as opposed to seeing it in isolation.
The refugee crisis and the rise in Euroscepticism, which were shown above to contribute towards a less one-dimensional political spectrum, are also shorter-term issues that feed into a sense of crisis.The short-term trends of the Financial Crisis, the refugee crisis and Euroscepticism, combined with the long-term trend of the exponential increase in globalisation of the past 30 years (Inglehart, 1990) may therefore have contributed to the current peak in populism.

Historical Factors
As mentioned above, this article assumes that the median voter theorem never holds completely, but that it breaks down more significantly at times when there is a peak in populism.The perspective offered is that of recurrent break-downs in the theorem, leading to peaks in populism.
The historical, cyclical and temporally dynamic nature of populism is therefore important.The cyclicality of populism has been extensively studied for the case of Latin America's left-wing populism (see, for example, Dornbusch and Edwards, 1991).However, there is a dearth of research on the cyclicality and temporal dynamics of right-wing and non-partisan populism, which is currently prevalent in the United States and Europe.
While there are studies focusing on the cyclicality of the support for individual populist parties (see, for example, Anderson, 1996, for Denmark and Norway), comparative studies on the cyclicality of populism worldwide, and in its various forms, are lacking8 .This article does not aim to offer such a study, but it does aim to contribute towards the study of cyclicality and temporal dynamics of populism, by basing the analysis of the current peak in populism on the view that it is in great part caused by the convergence of cyclical trends in Europe and the United States.
It is argued that populism, while always an element of democratic systems (Taggart, 2004), does become more acute at certain periods in time.I argue that this evolution can most usefully be seen as a succession of peaks and troughs, and that it is possible to detect a predictable pattern similar to that of business cycles in economics.In the United States, it is possible to identify regular peaks of populism every 20 or 30 years (see, for example, Kazin, 1998).Those are visible in the 1890s (the People's Party, pitting small farmers against industrial interests), the early 20 th Century (socialist and prohibitionist movements), the 1940s (anti-Communism and resurgent Conservatism), the 1960s (George Wallace and the white backlash), the 1980s (Reagan and the Christian Right) and the 2010s (rise of the Tea Party movement, followed by the election of Donald Trump in 2016).
In Europe, the dynamics have been very different, particularly as the post-War period saw comparatively little populism (Taggart, 2004) and was significantly impacted by the historical upheavals of two world wars, followed by a Cold War.Following the well documented peak in populism of the 1930s, involving the rise of Hitler and Mussolini, and culminating with the Second World War, populism became less attractive (Rydgren, 2005).From the 1940s until the end of the 20 th century, populism in Europe was less prevalent than it is today (Mudde, 2004), with the traumatic effect of two wars, combined with the ideological complications of the Cold War that followed, leading to a distrust of populism in both Western and Eastern Europe.Rydgren (2005) argues similarly that the stigmatization of Nazism and anything associated with it led to a fall in attraction for populism in Europe from the 1940s to the 1970s, boosted by the strong post-war economic development in Europe.While populist movements remained present throughout those years9 , the literature agrees that they were less prevalent than in the current times (see, for example, Mudde, 2004, Taggart, 2004), and it is difficult to identify any peak during this period. 10With memories of the (real and cold) wars fading with the end of the 20 th century, the 21 st century is seeing the re-emergence of populism.This trend first became a concern in Eastern Europe (Bugaric, 2008), but has now become significant in Western Europe as well.Therefore, a cyclical peak of populism in the United States is coinciding with both a culmination of the gradual rise of populism in Europe since World War 2, and a smaller peak in populism in Europe caused by recent events, including the refugee crisis and the aftermath of the 2008 Great Recession.This coincidence between a European peak and an American peak contributes to the perfect storm of populism.

Assumption 4: Parties can move freely within the spectrum
"4.Once placed on the political scale, a party can move ideologically either to the left or to the right up but not beyond the nearest party towards which it is moving."(Downs, 1957, p. 142) All the assumptions so far focused on the demand-side (the voters); this assumption now focuses on the supply-side (the politicians and political parties).This assumption implies that political parties will be able to move within the spectrum freely, as to position themselves in a point where they can maximise their potential number of votes.In other words, it implies that the supply-side is functioning well in the democratic system, and that political parties are able to offer policies that match the demand from voters.If that assumption fails, voters will not be supplied with a good variety of policies to choose from, and this may explain the attraction to populism.
To assess whether this is happening now, it is useful to extend the analogy with market supply and demand, to find out whether the variables that can affect supply in a market (cost, price, demand, level of competition and behaviour of producers) can explain the current situation, either through a higher supply of populism, or, indirectly, through a dysfunction of the supply side which can be used by populists in their favour11 .The previous section about assumption 3 has already shown that there was an increase in demand, so this section will focus on the other factors.
It can be argued that the "cost" of supplying populism has been lowered by changes in the media, and in particular by the prevalence of social media.New forms of media are more affordable, making it easier for politicians who don't have the financial backing of mainstream parties to catch the attention and interest of voters (Hong, 2013).The use of social media has been an important factor in the rise of populist candidates and movements, as evidenced most notably by Donald Trump's use of Twitter.
The "price" can be said to change if there is an alteration in the benefit that suppliers derive from an extra voter.This happens if there is a greater opportunity to attain office, for example than a populist movement.because populist parties which before would not have had that chance are now able to have it, as populism has become more acceptable and "electable".The past years have indeed witnessed a greater electability of populist politicians; for example, Donald Trump's 2000 campaign was unsuccessful, while the same politician achieved victory in 2016 (Oliver and Rahn, 2016).It can therefore be concluded that the price of populism has risen, leading to higher supply.
Another factor that be said to differ is the behaviour of "suppliers".Under perfect market conditions, suppliers need to have as their main objective the maximisation of profit -or, in this case, votes.However, populist parties will often not be vote-seeking parties, but rather policyseeking parties (Strom, 1990), which are not aiming to maximise their votes but instead to have an influence on policymaking.As noted in the above paragraph, several populist movements and politicians were seen as unelectable until recently, and indeed many still have not achieved office, even though the current wave of populism has provided them with the opportunity to reach objectives far beyond what was within their reach beforehand.Starting with objectives other than vote-maximisation leads to populist politicians positioning themselves elsewhere than at the median.Indeed, extreme-left (such as Syriza or Podemos) or extreme-right (such as Front National or Alternatief fuer Deutschland) positions are often present.This, together with voters' interest in such extreme positions, contributes towards a malfunction of the median voter theorem.
More generally, populists can also benefit if there is a dysnfunction in the supply side, which will make them more attractive to voters who feel that they are not being adequately represented.Oliver and Rahn (2016) show that Trump and other populist successes in the US were caused by a representation gap, whereby mainstream parties were not adequately representing the views of the voters.Mainstream parties struggle to address divisive issues such as immigration and EU integration, leaving a vacuum in certain spots of those spectrums which can be taken by populists.
7. Assumption 5: Extremist voters may abstain if they perceive the main parties as too similar "5.In a two-party system, if either party moves away from the extreme nearest it toward the other party, extremist voters at its end of the scale may abstain because they see no significant difference between the choices offered them." (Downs, 1957, p. 142) This assumption foresees that extremist voters will abstain because their extreme views are not represented by any of the parties with a realistic chance of winning.This assumption can be said to break down when extremist voters and extreme views, which are normally not represented in the political system, start to be represented.Indeed, populist movements often include the political mobilisation of sections of the electorate which would otherwise have felt alienated by the political system (Oliver and Rahn, 2016).Extreme views are therefore legitimised and become part of the political discourse.These include views that would normally be seen as xenophobic (such as an opposition to immigration or targeting specific origins by religion or ethnicity), paranoid (such as believing in a world-wide conspiracy between banks and large companies) or unviable (such as complete opposition to capitalism).I argue that this legitimisation is happening now more than usual, because the current populist climate has led to views which are normally seen as unacceptable being legitimised.This is a self-fulfilling prophecy: as populism rises, populist views become more acceptable, which in turn fosters more populism (Mudde, 2004).The literature on "contagion" effects of populism (Roduijn et al, 2014, Rydgren, 2005) is useful to explain how populist gains in one party/ country can lead to populist gains somewhere else.However, so far the term "contagion" in the populism literature has been used in relation to programmatic contagion, this happening when mainstream parties are influenced by populist parties (see, for example, Roduijn et al, 2014).Instead, this article is concerned with the international contagion of populism itself, or the fact that a peak of populism in one country can spread to other countries.
A case in point occurred in January 2017, when the leaders of several European populist parties met in Germany to celebrate the election of Donald Trump, and told the press that Trump's victory would spur their own parties (Connolly, 2017).Rydgren (2005) studies the contagion between right-wing populist parties in Europe, and argues that those parties mutually legitimize each other, in particular through the creation of an ideological master frame based on nativism and authoritarianism.I argue that this contagion effect goes beyond Europe (as can be seen from the above mentioned references to Trump by European populist parties) and is extended by the increasing globalisation of media and exchange of ideas.It can therefore be argued that one of the forces contributing towards the current rise in populism is a snow-ball effect, whereby the gradual rise in populism has in turn led to a legitimization of extremist political views, which in turn promotes further gains in populism.

Conclusion
This article has proposed a theoretical framework for the analysis of populism, based on an alteration of the equilibrium state of the median voter theorem.It has shown that populism can be directly linked with basic assumptions related to the functioning of the political system.In an ideal world where those assumptions were fully met, the theorem would apply fully, and policymaking would fully meet the preferences of the median voter.At the other extreme, when several of those assumptions break down, this results in outcomes that appear to run counter the logic of democracy, such as the seemingly irrational voting for populist politicians, resulting in outcomes that clearly do not meet median, average or balanced preferences.
While none of the theorem's assumptions is ever fully met in the real world, this article found that all the assumptions were failing more than usual at present, and that this can be associated with the current peak in populism.In particular: the left-right spectrum has lost its relevance; the distribution of ideological preferences is not normal due to a polarisation of society; sociological, psychological and historical factors are affecting voters' preferences more than usual; political party competition has been disrupted; and extreme political preferences have been legitimised.
The combination of all these factors leads to the current perfect storm of populism.
The article builds on existing research on the causes of populism (for example, Inglehart and Norris, 2016), and extends it to explain the current peak in populism.Rather than attempting to find one key narrative to explain populism, this article takes a broader approach by looking at different elements of the political system, and assessing how each of them contributes to populism.In doing so, importance is given to the phenomena of contagion and cyclicality, and the role that they play in bringing together the different factors.For example, the way in which demand side changes feed into supply side changes is analysed, as well as how changes in one country carry through to another.
It is important to have a theoretical framework under which to model populism because this provides scholars with a tool to examine and compare this phenomenon, both between countries, and over time.Such a comparative approach makes it possible to understand why populism varies across time and regions, and to analyse what causes those variations.It also contributes towards understanding the ongoing peak in populism, which in turn helps analyse how can it be addressed optimally.
The main limitation of this study is that the model proposed, as with any model, is imperfect and needs to be used with caution to avoid over-simplification.The phenomena described are highly complex, and to model them inevitably involves making simplifying assumptions.It is therefore important to bear in mind that not all the complexities of the phenomena have been captured by the model, as indeed no model can fully reflect reality.In particular, populism is not the only reason why the median voter theorem's assumptions are challenged, and not all of those challenges result in populism.
This article uses the median voter theorem as a benchmark -a model that describes an unrealistic situation of perfect competition in the political system, against which the real world is measured.This contrasts with the usual approach to the theorem in the literature, whereby it is either assumed to hold perfectly, or rejected outright (Holcombe, 1989).This approach may be extended to other areas of research, and could be useful to analyse other aspects of the political system.
For example, if the median voter theorem is seen as the benchmark of perfect democratic system, deviations from that benchmark can also be useful to understand other issues that affect the democratic system, such as declining interest and participation in the political system by the voters, or changes in the media and how this affects politics.The wider definition of the theorem proposed in this paper makes it relevant to analyse not only the left-right political spectrum, but also any other issues which potentially destabilise the political system.