1.1 Crises of Democracy

What should we be looking for if we fear that democracy is experiencing a crisis? To identify crises of democracy, we need a conceptual apparatus: What is democracy? What is a crisis? Is the crisis already here or is it only impending? If it is already here, how do we recognize it? If it is not yet visible, from what signs do we read the future?

We are repeatedly told that “Unless democracy is X or generates X, …” The ellipsis is rarely spelled out, but it insinuates either that a particular system is not worthy of being called a “democracy” unless some X is present or that democracy will not endure unless some X is satisfied. The first claim is normative, even if it often hides as a definition. Skinner (Reference Skinner1973: 303), for example, thinks that a system in which only some people rule does not merit being called a “democracy,” even if it is a competitive oligarchy. Rosanvallon (Reference Rosanvallon2009), in turn, claims that “Now power is not considered fully democratic unless it is submitted to the tests of control and validation at the same time concurrent and complementary to the majoritarian expression.” The second claim is empirical, namely, that democracy may not endure unless some Xs are present (or absent). If democracy requires some conditions – say J.S. Mill’s (Reference John Stuart and Robson1977: 99) “high wages and universal reading” – just to function, then it is vulnerable to breakdowns when these conditions are absent. A modicum of economic welfare, some level of citizen’s confidence in political institutions, or some minimal level of public order are the most plausible candidates for such conditions.

Thus, one way to think is that democracy experiences a crisis when some features which we consider as definitional of democracy are absent. Consider a triad of what Ginsburg and Huq (Reference Ginsburg and Huq2018a) consider to be “the basic predicates of democracy”: competitive elections, liberal rights of speech and association, and the rule of law. If we treat this triad as definitional, we get a ready-made checklist of what we should be looking for to identify crises of democracy: elections that are not competitive, violations of rights, breakdowns of the rule of law. Yet if we believe that democracy may not survive given some particular situation, we may still be worried that it faces a crisis even if no such violations are observed. We may still have a checklist constructed by the definition but now we also have a set of hypotheses that condition the survival of democracy on some potential threats, and we are directed by these hypotheses to examine the particular threats. If such hypotheses are valid, if the survival of democracy depends on some aspects of its performance, and democracy does not generate the required outcomes, a crisis occurs – democracy is in crisis.

Note that some features may be treated alternatively as definitional or as empirical. If one defines democracy as Rosanvallon does, to include contramajoritarian constraints on majority rule, “constitutional democracy,” then the erosion of judicial independence is prima facie evidence that something is wrong. But one may also reason that if the judiciary is not independent, the government will be free to do whatever it wants, violate the liberal right, or make elections non-competitive. The problem with adding adjectives to “democracy” is that not all good things must go together. The more features – “electoral,” “liberal,” “constitutional,” “representative,” “social” – we add to the definition of democracy, the longer the checklist, and the more crises we will discover. In contrast, the same list can be treated as a set of empirical hypotheses. We can then investigate empirically what are the conditions for elections to be competitive or for rights to be observed or for the rule of law to prevail. If it is true that elections are competitive only if rights are observed and law rules, then taking any one of these features as definitional and treating others as “preconditions” is coextensive. If they are not coextensive, then some kind of definitional minimalism is unavoidable: we must choose one of the potential features as definitional and treat others as hypothetical conditions under which the selected feature is satisfied.

Hence, what we would consider as crises and how we should go about diagnosing them depends on how we think about democracy. The view of democracy I adopt is “minimalist” and “electoralist”: democracy is a political arrangement in which people select governments through elections and have a reasonable possibility of removing incumbent governments they do not like (authors who held this view include Schumpeter Reference Schumpeter1942, Popper Reference Popper1962, and Bobbio Reference Bobbio1987). Democracy is simply a system in which incumbents lose elections and leave when they lose. Hence, I investigate the possible threats to elections becoming non-competitive or inconsequential for whoever remains in power. To repeat, these threats may include violations of the preconditions for contested elections enumerated by Dahl (Reference Dahl1971) – the liberal rights and the freedoms – simply because without them the incumbent government could not be defeated. They may also include breakdowns of the rule of law and erosion of the independent power of the judiciary, along with loss of confidence in representative institutions (as in “representative democracy”), acute inequality (as in “social democracy”), or the use of repression to maintain public order (“liberal democracy”). But I treat these violations as potential threats to the ability of citizens to remove governments by elections, not as definitional features of “democracy.”

The relation between “democracy” in the minimalist sense and the “rule of law” is particularly complex. First, there are both logical and empirical reasons to question whether supra-majoritarian institutions, such as bicameralism or presidential veto, or counter-majoritarian institutions, such as constitutional courts or independent central banks, are necessary to support the rule of law. Gargarella (Reference Gargarella, Maravall and Przeworski2003), for example, lists several mechanisms by which a majority can and would want to constrain itself even in the absence of such institutions. As McGann (Reference McGann2006) observes, there are well-established democracies, including the United Kingdom and Sweden, which have neither a separation of powers nor judicial review of the constitution, and yet in which majorities constrain themselves from violating rights. Indeed, Dixit, Grossman, and Gull (Reference Dixit, Grossman and Gul2000: 533) demonstrate logically that violations of rights are likely to be more egregious in the presence of supra-majoritarian institutions once a government enjoys supra-majority support.

Second, I put “rule of law” in quotation marks because, as Sanchez-Cuenca (Reference Sanchez-Cuenca, Maravall and Przeworski2003: 62) astutely put it, “The law cannot rule. Ruling is an activity, and laws cannot act.” What is typically seen as a relation between democracy and the rule of law is in fact a relation between populated institutions: governments and courts (Ferejohn and Pasquino Reference Ferejohn, Pasquino, Maravall and Przeworski2003). Law “rules” when politicians and bureaucrats obey judges, and whether politicians do or do not comply with the instructions of constitutional justices is a contingent outcome of their electoral incentives. Moreover, as will be seen below, it is often next to impossible to determine if some particular measures they adopt do or do not conform to legal or constitutional norms, with individual judgments, including those of constitutional justices, clouded by partisanship. Under democracy, the only effective device for disciplining politicians are elections: as Dixit, Grossman, and Gull (Reference Dixit, Grossman and Gul2000: 533) observe, “The ruling individuals must foresee an appreciable chance that their power will come to an end … And they must foresee a possibility of regaining power once it is lost.” There are two possibilities: (1) politicians (and bureaucrats) obey judges because otherwise they would lose elections, so that “the law” rules; (2) politicians do not obey judges because otherwise they would lose elections – a majority does not want politicians to listen to what the judges tell them they can or cannot do. The rule of law is violated but as long as politicians’ actions are motivated by the fear of losing elections, the system is still democratic by the minimalist criterion. Democracy is “illiberal” – a term made fashionable by Zakaria (Reference Zakaria1997) and embraced by the Hungarian Prime Minister, Viktor Orbán – but it is illiberal because politicians expect that otherwise they would lose elections. Yet, if politicians do not obey the judges even if a majority would want them to because they do not fear elections, the regime is not democratic.

Understood in this way, democracy is a mechanism for processing conflicts. Political institutions manage conflicts in an orderly way by structuring the way social antagonisms are organized politically, absorbing whatever conflicts may threaten public order, and regulating them according to some rules. An institutional order prevails if only those political forces that have institutionally constituted access to the representative system engage in political activities, and if these organizations have incentives to pursue their interests through the institutions and incentives to temporarily tolerate unfavorable outcomes. Specifically, conflicts are orderly if all political forces expect that they may achieve something, at the present or at least in some not too distant future, by processing their interests within the institutional framework while they see little to be gained by actions outside the institutional realm. Hence, democracy works well when whatever the conflicts that arise in society are channeled into and processed through the institutional framework – most importantly elections, but also collective bargaining systems, courts, and public bureaucracies – without preventing anyone from gaining access to these institutions just because of the substance of their demands. To put it succinctly, democracy works when political conflicts are processed in liberty and civil peace.

The conflicts that divide a particular society at a particular time may be more or less intense and may divide the society along different lines depending on whether they concern economic interests, cultural values, symbolic issues, or just fleeting passions. Their forms, their subjects, and their intensity depend on the actions of governments and the alternatives offered by competing political forces. The stakes entailed in institutionalized conflicts do not simply reflect the intensity of antagonisms that arise in a society. Institutional frameworks shape the ways in which social conflicts become politically organized, some increasing and others limiting the stakes in the outcomes of political competition. I argue below (see Chapter 9) that democracy works well when the stakes entailed in institutionalized conflicts are neither too small or too large (for a technical version of this argument, see Przeworski, Rivero, and Xi Reference Przeworski, Rivero and Tianyang2015). The stakes are too low when results of elections have no consequences for people’s lives. They are too high when results of elections inflict intolerable costs on the losers. When people believe that results of elections do not make a difference in their lives, they turn against “das System,” as in Weimar Germany. When the electoral losers discover that the government pursues policies that significantly hurt their interests or values, they become willing to resist the government by all – including violent – means, as did the bourgeoisie in Chile under President Allende. Hence, democracy works when something is at stake in elections but not too much is at stake.

An often overlooked emphasis of Schumpeter’s (Reference Schumpeter1942: chapter 23, section 2) “minimalist” view of democracy is that governments must be able to govern and must govern competently. Later I delve into some historical periods in which the institutional framework made it difficult for governments to be able to govern, either because the electoral system led to government instability, as in Weimar Germany and the French Fourth Republic, or because the system of separation of powers generated a stalemate between the executive and the legislature, as in Allende’s Chile. To govern effectively, governments must satisfy a majority yet not ignore the views of intense minorities. When conflicts are intense and a society is highly polarized, finding policies acceptable to all major political forces is difficult and may be impossible. There are limits to what even the best-intentioned and competent governments can do.

If this is the standard, when is democracy “in crisis”? The very word “crisis” originates from ancient Greek, where it meant “decision.” Crises are situations that cannot last, in which something must be decided. They emerge when the status quo is untenable and nothing has yet replaced it. This is what we mean when we say that “the situation reached a crisis point”: when doctors say someone is in a crisis, they mean that the patient will either recover or die but cannot remain in the current state. Crises may be more or less acute: in some a turning point may be imminent but some crises may linger indefinitely, with all the morbid symptoms.

The intuition of crises conveyed by Gramsci’s motto is that the current situation is in some ways untenable, that some threat to democracy has already materialized, yet the status quo democratic institutions remain in place. While Marx (Reference Marx1979 [1859]: 43–4) thought that “new superior relations of production never replace older ones before the material conditions for their existence have matured within the framework of the old society,” nothing guarantees that when the status quo institutions malfunction, some other institution would descend on earth as a deux ex machina. What happens when the status quo institutions do not generate desirable outcomes depends on their properties and on the alternative institutions – would any do better? – on exogenous conditions, and on the actions of the relevant political forces under these conditions. That a disaster is unfolding under the status quo institutions need not imply that some other institutions would do better: this was Winston Churchill’s view of democracy. But even if some alternatives are feasible, it may well be that given the relations of political power under the extant institutions, the situation would linger on and on. Crises are then situations in which the condition under the status quo institutions is some kind of a disaster: no change occurs, but it may. This is what we will be looking for below: whether the current situation is in some ways threatening and whether there are signs that the traditional representative institutions are being affected.

“Crises of capitalism” deserve a separate comment. Capitalism – an institution that combines private ownership of most productive resources with the allocation of resources and distribution of incomes by markets – periodically generates “crises,” understood as periods in which incomes fall sharply and either inflation flares or unemployment soars or both, as during the “stagflation crisis” of the 1970s, a combination of high inflation with high unemployment caused by a jump in prices of raw materials (Bruno and Sachs Reference Bruno and Sachs1985). But are economic crises “crises of capitalism”? They would be if one expects that when the economy is in the doldrums, capitalism will or at least may collapse. But an implosion of capitalism is not in the realm of the possible. When a famous leftist economist, Michal Kalecki (Reference Kalecki1972 [1932]), asked in 1932, at the worst moment of the Great Depression, “Is a capitalist exit from the crisis possible?,” his argument was that, even if the adjustments required to exit from economic crises are painful and may take time, capitalism is a self-correcting system. Prices and wages may be sticky but eventually supply and demand adjust, the crisis is over, and capitalism is still here. It can be abolished by a political revolution – a possibility Kalecki did entertain and Communists implemented – but not implode. The general lesson for understanding crises is that some institutions are impervious to the outcomes they generate, so that crises which occur under them do not turn into crises of the institutions.

Disasters that occur under democracy, however, may turn into crises of democracy. Borrowing their list from Habermas (Reference Habermas1973: 49), disasters are situations in which

• – the economic system does not produce the requisite quantity of consumable values, or;

• – the administrative system does not produce the requisite quantity of rational decisions, or;

• – the legitimation system does not provide the requisite quantity of generalized motivations, or;

• – the socio-cultural system does not generate the requisite quantity of action-motivating meaning.

This list, however, is too abstract to guide research. The observable candidates for disasters are economic crises, intense conflicts in society, and political paralyses, situations in which the government is unable to govern given the particular form of democratic institutions.

When we think that the situation is in some way threatening, we look for signals – harbingers of change. Several countries, ranging from Canada in 1931–3 to Uruguay in 2001–3, experienced profound economic crises with almost no political repercussions and no signals of democracy being weakened. Yet in some situations crises in other realms – whether economic, cultural, or autonomously political (say corruption scandals, as in Italy in 1993 or in Brazil now) – manifestly weaken the established democratic institutions. The visible signals that democracy is in crisis include a sudden loss of support for established parties, withdrawal of popular confidence in democratic institutions and politicians, overt conflicts over democratic institutions, or an incapacity of governments to maintain public order without repression. Perhaps the most tangible sign of a crisis is a breakdown of public order: in the words of Linz (Reference Linz1978: 54), “The most serious crises are those in which the maintenance of public order becomes impossible within a democratic framework.” Democracy is in crisis when fists, stones, or bullets replace ballots. Either the incumbents make it impossible for the opposition to remove them from office and the opposition has no other avenues than resistance, or the opposition does not recognize the legitimacy of the government and the government defends itself by repression, or antagonistic political groups do not accept the outcomes of the institutional interplay of interests and revert to direct, often violent, confrontations. When such situations extend over time, public order breaks down, everyday life becomes paralyzed, and violence tends to spiral. Such crises become mortal when the design of democratic institutions generates institutional stalemates, as in Weimar Germany or in Chile under President Allende.

Institutions may generate outcomes that are intolerable for some and wonderful for others. Moreover, people may differ in their normative attachments: some valuing liberty more than order, others being willing to sacrifice it for the promise that trains would run on time (Mussolini promised they would under fascism, but they did not). Hence, to understand crises it is necessary to think in terms of conflicting interests and values. The poor are dissatisfied when their incomes stagnate, the rich enjoy their wealth and power, while some people, whether poor or wealthy, may care about political and economic inequality per se. Solutions to crises are likely to be controversial and subject to political conflicts. They depend on what the relevant political actors do under the circumstances. To this extent, therefore, they are indeterminate ex ante. Will a reduction of economic inequalities restore the political vitality of democracy? Will restrictions on immigration appease radical Right sentiments? Will some tinkering with representative institutions restore confidence in these institutions? Because the actors in crisis may choose different courses of actions, with different consequences, the best we can strive to determine is what is and what is not possible, perhaps with some cavalier forecasts about what is most likely.

What, then, are the possible outcomes of crises? Not all crises are mortal: some end in restoring the status quo ante, a return to “normalcy.” The sources of a crisis sometimes conveniently disappear. Democracy may be in a crisis when society experiences an economic disaster, but the crisis may dissipate when prosperity returns. Some crises can be overcome by partial reforms. The group that benefits under extant institutions can make concessions to the groups that suffer most under them. Such concessions have to be credible, because otherwise these groups will expect that they would be withdrawn once the crisis is over. Hence, concessions must entail some institutional reforms: the classical example is the extension of suffrage to the lower classes, which neutralized the threat of revolution by changing the income location of the decisive voter (Acemoglu and Robinson Reference Acemoglu and Robinson2000). Yet when we think about democracy what we fear is the prospect that some political forces would successfully claim that the only way to remedy some already occurring disasters – economic crises, deep-rooted divisions in society, breakdown of public order – is to abandon political liberty, unite under a strong leader, and repress pluralism of opinions, in short autocracy, authoritarianism, or dictatorship, whatever one wants to call it. The impending cataclysm is that democracy would either collapse outright or gradually erode beyond the point of no return.

The specter that haunts us today, I believe, is the last possibility: a gradual, almost imperceptible, erosion of democratic institutions and norms, subversion of democracy by stealth, “the use of legal mechanisms that exist in regimes with favorable democratic credentials for anti-democratic ends” (Varol Reference Varol2015). Without manifest signs that democracy has broken down, the line becomes thin, as evidenced by labels such as “electoral authoritarianism” (Schedler Reference Schedler2006), “competitive authoritarianism” (Levitsky and Way Reference Levitsky and Way2010), “illiberal democracy” (Zakaria Reference Zakaria1997), or “hybrid regimes” (Karl Reference Karl1995, Diamond Reference Larry2002). “Backsliding,” “deconsolidation,” or “retrogression” need not entail violations of constitutionality and yet gradually destroy democratic institutions.

To summarize this concept of “crisis of democracy,” think schematically as follows. Given some exogenous shocks, democracy generates some outcomes, positively or negatively evaluated by people with heterogeneous preferences over these outcomes and over the democratic institutions per se. Outcomes that threaten the continued existence of the traditional democratic institutions constitute “disasters.” Whether a particular situation qualifies as a crisis must be read from some manifest signals that democratic institutions are under threat. We are attentive to such signals because they may constitute harbingers of democratic collapse or gradual erosion. Yet the potential solutions to crises may include restoration of the institutional status quo, some partial reforms of traditional representative institutions that still preserve democracy, as well as its either abrupt or gradual destruction.

Why would democracies be vulnerable to crises? One must not forget that democracy is but a speck of human history, recent and still rare. It was born only in 1788, when the first national-level election based on individual suffrage took place in the United States; the first time in history that the helm of the government changed as a result of an election was in 1801, also in the United States. Use of force – coups and civil wars – remained frequent: between 1788 and 2008 political power changed hands as a result of 544 elections and 577 coups. Electoral defeats of those in power were rare until very recently and peaceful changes of governments even less frequent: only about one in five national elections resulted in the defeat of incumbents and even fewer in a peaceful change in office. As of today, sixty-eight countries, including the two behemoths, China and Russia, have never experienced a change in office between parties as a result of an election. Democracy is a historical phenomenon. It developed under specific conditions. It survived in some countries as these conditions evolved, but can it survive under all conditions?

Two structural conditions, I think, deserve special attention. The first is that political equality, which democracy is supposed to be based on, coexists uneasily with capitalism, a system of economic inequality. The second is the sheer quest for political power, whether or not based on economic interests.

1.2 Democracy and Capitalism

The relation between democracy and capitalism is subject to contrasting views. One claims a natural affinity of “economic freedom” and “political freedom.” Economic freedom means that people can decide what to do with their property and their labor endowments. Political freedom means that they can publicize their opinions and participate in choosing how and by whom they will be governed. But equating the concepts of “freedom” in the two realms is just a play on words. Looking into history shows that we should be surprised by the coexistence of capitalism and democracy. In societies in which only some people enjoy productive property and in which incomes are unequally distributed by markets, political equality combined with majority rule presents a threat to property. Indeed, beginning with Henry Ireton’s speech in the franchise debate at Putney in 1647, almost everyone had thought that they could not coexist. The English conservative historian and politician Thomas Macaulay (Reference Macaulay1900: 263) vividly summarized in 1842 the danger presented to property by universal suffrage:

Nine years later, from the other extreme of the political spectrum, Karl Marx (Reference Marx1952: 62) expressed the same conviction that private property and universal suffrage are incompatible:

The combination of democracy and capitalism was thus for Marx an inherently unstable form of organization of society, “only the political form of revolution of bourgeois society and not its conservative form of life” (Reference Marx1934 [1852]: 18), “only a spasmodic, exceptional state of things … impossible as the normal form of society” (Reference Marx and Draper1971 [1872]: 198).

These dire predictions turned out to be false. In some – specifically thirteen countries – democracy and capitalism coexisted without interruptions for at least a century, and in many other countries for shorter but nevertheless extended periods, most of which continue today. Working-class parties that had hoped to abolish the private property of productive resources realized that this goal is unfeasible, and learned to value democracy and to administer capitalist economies whenever elections brought them into office. Trade unions, also originally viewed as a mortal threat to capitalism, learned to moderate their demands. The outcome was a compromise: working-class parties and trade unions consented to capitalism, while bourgeois political parties and organizations of employers accepted some redistribution of income. Governments learned to organize this compromise: regulate working conditions, develop social insurance programs, and equalize opportunities, while promoting investment and counteracting economic cycles (Przeworski Reference Przeworski1986).

Yet perhaps this compromise is now broken. Unions lost much of their capacity to organize and discipline workers and with it their monopoly power. Socialist parties lost their class roots and with them their ideological as well as policy distinctiveness. The most visible effect of these changes is the sharp decline in the share of incomes from employment in the value added and, at least in the Anglo-Saxon countries, a steep increase of income inequality. Combined with a slowdown of growth, rising inequality causes many incomes to stagnate and income mobility to decline.

Is the coexistence of democracy and capitalism conditional on a continual improvement of material conditions of broad sectors of the population, either because of growth or because of increasing equality? History indicates that democracies are solidly entrenched in economically developed countries and impervious to economic as well as other crises, even of a large magnitude. But is history a reliable guide to the future?

1.3 Democracy and the Quest for Power

The second reason democracies may experience crises is inherent in political competition. The dream of all politicians is to conquer power and to hold on to it forever. It is unreasonable to expect that competing parties would abstain from doing whatever they can do to enhance their electoral advantage, and incumbents have all kinds of instruments to defend themselves from the voice of the people. They are able to consolidate their advantage because they constitute a legislative majority and because they direct public bureaucracies. Although at times they are constrained by independent courts, control over legislation grants incumbents an opportunity to adopt legal regulation in their favor: just think of voter registration, manipulation of electoral systems, or gerrymandering. The courts or some other independent bodies may invalidate some such attempts but not always have reasons or the will to do so: there are many ways to carve districts, each with electoral consequences, which are not blatantly discriminatory. In turn, as principals of ostensibly non-partisan bureaucracies, incumbents can instrumentalize them for partisan purposes. Control over the apparatuses of repression plays a particularly important role in undermining all or some opposition. Exchange of favors for financial resources is yet another source of advantage. And, when all else fails, fraud is the last resort.

The question is why some political leaders use these methods while others are content with letting the people decide and being willing to leave office when people do so decide. Their motives matter and so do the constraints. When political parties are highly ideological, when they believe that essential issues or values are at stake, they see their opponents as enemies who must be prevented from coming to office by any means. In Poland the ruling party, PiS (Law and Justice), believes that the very values that constitute Poland as a Christian Nation are at stake and all their opponents are “traitors.” In Hungary, President Orbán thinks that what is at stake is whether “Europe will remain the continent for Europeans.” Hence, both attempt to control the media, restrict freedom of association, pack state agencies with their partisan supporters, and toy with electoral rules. These actions are intended to relax the electoral constraints they face, and to make an electoral victory of the opposition next to impossible. Yet they still face political, rather than narrowly electoral, constraint: various forms of popular resistance, such as mass demonstrations, political strikes, or riots. They face the danger that political conflicts could spill out of institutional bounds, resulting in a breakdown of public order. They may or may not take this risk, and if they do, democracy is in crisis.

1.4 A Preview

How then should we go about determining if democracy is presently in crisis, or at least if a crisis is impending?

To look into the future, to identify the possibilities latent in the current situation, we first need to see if we can learn something from the past. Under what conditions did democratic institutions fail to absorb and peacefully regulate conflicts? To answer this question, Part I summarizes the historical experience of all democracies that have been at one time or another consolidated, in the sense of having experienced at least two peaceful alternations in office that resulted from elections, comparing some observable conditions of the democracies that fell and those that survived. Such comparisons, however, are inevitably static, while the outcomes that emerge under any conditions are highly contingent, depending on who does what when. To develop intuitions, I delve in more detail into four cases: the Weimar Republic between 1928 and 1933, and Chile between 1970 and 1973, are two flagrant instances in which democracy succumbed, while France and the United States in the 1960s are cases of political repression and breakdown of order that were resolved institutionally.

Yet history does not speak for itself. Can we trust its lessons? Lessons from history are relatively reliable when current conditions imitate those observed in some past, but iffy when they are unprecedented (King and Zheng Reference King and Zheng2007). Hence, to see if history can be our guide, we need to compare the current situation with those of the past. Do the current conditions resemble those of democracies that fell or of those that survived? Or are they unprecedented? Some aspects of the current situation are new, in particular a rapid destabilization of traditional party systems. So is the stagnation of low incomes as well as the erosion of the belief in material progress. But causal links are far from obvious. Is the current political conjuncture driven by economic trends or by cultural transformations, or is it autonomous from changes in the economy and society? At what level should we seek explanations: general trends, such as globalization, or specific situations of particular individuals, say those who fear losing decently paying jobs? These are the questions considered in Part II.

To assess the prospects for the future, we need to understand how democracy works when it works well, which is the subject of the theoretical chapter that opens Part III. With this understanding, we can consider the foreboding and uncharted possibility of a gradual erosion of democracy, its subversion by elected governments. Finally, even if we cannot tell what is most likely to ensue, we can at least speculate about what is and what is not possible. Can it happen here?