Book Summary: How Democracies Die by Steven Levitsky and Daniel Ziblatt

Book Cover for How Democracies Die by Steven Levitsky and Daniel Ziblatt

This summary of How Democracies Die: What History Reveals About Our Future by Steven Levitsky and Daniel Ziblatt explains how American democracy is more fragile than we may think and suggests ways to prevent it from dying.

Buy How Democracies Die at: Amazon | Kobo (affiliate links)

Key Takeaways from How Democracies Die

  • Democracy is hard work, and can be frustrating for all involved.
    • Outsiders may find democracy’s checks and balances too frustrating and seek to bypass them. Elected leaders can kill democracies, by undermining the very systems that elected them.
    • The authors suggest four warning signs to help identify authoritarian leaders — and Donald Trump met every one of them even before he became president.
    • But democracy in America has been weakening for decades, long before Trump.
  • The first test for democracy is how well it keeps dangerous demagogues out of power.
    • America’s party establishment used to be quite good at keeping populist outsiders such as Henry Ford out of power.
    • But in 1972, the presidential primary system became the main method of selecting party nominees, which reduced party leaders’ gatekeeping role.
    • America failed this test when Donald Trump became the Republican nominee, and then President, in 2016.
  • The second test is the strength of unwritten democratic norms once an autocrat has come to power. In particular:
    • Mutual toleration — the idea that you see your political opponents as legitimate rivals. Unfortunately, today in America people increasingly see their opponents as “evil”.
    • Institutional forbearance — when leaders and officials exercise restraint in their powers. For example, not stacking the Supreme Court.
    • Mutual toleration and institutional forbearance are linked. If you see your opponents are evil, you’re less likely to exercise forbearance and more likely to use authoritarian measures to defeat them.
    • Both of these have been declining in the US since around the 1990s.
  • Once an autocrat comes into power, they can consolidate their power by:
    • Capturing the referees — judges, law enforcement and regulatory agencies;
    • Sidelining the opposition — political opponents, media, influential public figures, religious leaders;
    • Tilting the playing field — changing the constitution or electoral system to lock in their advantage; and
    • Exploiting a crisis — the courts and the public are more likely to tolerate authoritarian measures during a crisis.
  • So what can we do?
    • Don’t fight dirty.
    • Don’t eschew diversity.
    • Build a broad pro-democratic coalition.
    • Reform the Republican Party.
    • Focus on depolarising economic policies — i.e. universalistic policies over means-testing.

Detailed Summary of How Democracies Die

Democracy is hard

Democracy is grinding work, with many setbacks. It requires negotiation, compromise, and concessions. Victories always only partial.

Democracy can be frustrating for everyone, but democratic politicians know they have to accept them. Outsiders, on the other hand, may find democracy and its checks and balances too frustrating and seek to bypass them.

Elected leaders can kill democracies

When we think of democracies dying, we usually think of violent, military coups, such as in Chile, Argentina, Brazil, Thailand, Türkiye and Egypt.

But democracies can also die at the hands of elected leaders. After gaining power, leaders can dismantle the very democratic processes that brought them to power in the first place. They can do this gradually and even legally. It’s far more subtle than a military coup, and the autocrats may maintain a “veneer” of democracy. Some people do not even realise what is happening, and continue to believe they’re in a democracy. Since the end of the Cold War, this has been the more common way in which democracies die.

Four authoritarian warning signs

It’s not always clear whether a politician will become an authoritarian once in power. The authors suggest four warning signs. Authoritarians tend to:

  1. Reject, in words or action, the democratic rules of the game.
  2. Deny the legitimacy of their opponents.
  3. Tolerate or encourage violence.
  4. Show willingness to use power to curtail the civil liberties of opponents and critics.

Apart from Richard Nixon, no major-party presidential candidate before Trump met even one of these four warning signs over the last century. Trump met all four of these before becoming President in 2016:

  1. Trump questioned the legitimacy of the electoral process and suggested that he might not accept the results of the 2016 election. This was unprecedented.
  2. Trump vocally challenged the legitimacy of Barack Obama’s presidency by suggesting that he was born in Kenya and that he was a Muslim. He also denied Hillary Clinton’s legitimacy by calling her a “criminal” and saying that she “has to go to jail.”
  3. During the campaign, Trump encouraged his supporters to engage in violence against protestors.
  4. Trump repeatedly threatened to punish unfriendly media.

But American democracy has been weakening long before Trump

The authors point to two things that have made America more vulnerable to authoritarianism:

  • Establishment gatekeepers now have less power; and
  • Unwritten democratic norms, namely mutual toleration and forbearance, have gradually eroded (while partisanship has increased).

Establishment gatekeepers now have less power

Populists are anti-establishment politicians who claim to represent the voice of “the people”. They tend to deny the legitimacy of established parties and attack them as undemocratic. Extremists and populists will emerge from time to time, even in healthy democracies.

Established political parties are democracy’s gatekeepers. The party leaders can play a role in keeping authoritarians away from power in multiple ways:

  • They can keep would-be authoritarians off party ballots at election time;
  • Parties can root out extremists in the grassroots of their own supporters;
  • Instead of allying with or legitimising antidemocratic parties and candidates, they can isolate them; and
  • Prodemocratic parties may form a united front and join up with their ideological opponents to preserve democracy.
in extraordinary times, courageous party leadership means putting democracy and country before party.
— Levitsky and Ziblatt in How Democracies Die

Gatekeeping is admittedly undemocratic

There is an inherent tension between gatekeeping and openness. Nobody likes smoke-filled rooms, where a small group of elites who are not accountable to the public select the candidates. The authors acknowledge that overreliance on gatekeeping is undemocratic, and that smoke-filled rooms did not always produce good presidents.

However, they argue that overreliance on the “will of the people” can also be dangerous if people elect a demagogue who threatens democracy itself.

One oft-forgotten virtue of smoke-filled rooms was that they tended to keep demonstrably unfit figures off the ballot, as party elites opted for “safe” candidates instead. When candidates are not restrained by party allegiances, they have little to lose by stirring up mass hatred or making absurd promises. [Interesting. In this way, it seems parties force candidates to think longer-term, beyond their immediate election. ]

Examples where gatekeepers have kept populists out

The authors point to various examples, such as in Belgium, Austria and France, where establishment gatekeepers kept demagogues out of power.

Example: Henry Ford

Henry Ford, the founder of the Ford Motor Company, was one of the richest men in the world in the early 20th century. He was widely admired, especially in the Midwest, because of his “rags to riches” story.

Ford was also an extremist who railed against bankers, Jews, and Bolsheviks. (Adolf Hitler even wrote positively of him in Mein Kampf.) He flirted with the idea of running for president in 1924, but party leaders soundly rejected him. Despite his widespread popularity, party leaders effectively locked him out of contention as he had no training or experience for the role.

Examples where gatekeepers have allied with populists

The temptation for political elites to align with a popular outsider is stronger when they’re facing severe challenges, such as an economic crisis or waning popularity. These alliances lends outsiders enough respectability to become legitimate contenders for power.

Often the elites mistakenly believed they could control the outsider. Examples include Adolf Hitler, Benito Mussolini, and Hugo Chávez.

I have just committed the greatest stupidity of my life; I have allied myself with the greatest demagogue in world history.
— Alfred Hugenberg, a prominent conservative who helped Hitler, the day after Hitler became chancellor (as quoted by Levitsky and Ziblatt in How Democracies Die).

Example: Hugo Chávez

In the 1970s, Venezuela was viewed as a model democracy. However, its economy was dependent on oil. In the 1980s, the price of oil crashed and Venezuela’s economy entered into a slump for over a decade, Venezuelans grew disaffected. There were riots.

At this time, ex-president Rafael Caldera’s political career was waning, but he still dreamed of returning to the presidency. Caldera saw an opportunity to align with Chávez to get a popularity boost. Chávez was a political outsider who had mounted two failed coup attempts. In 1993, he was in jail awaiting trial for treason.

During Caldera’s electoral campaign, he promised to free Chávez, which was a popular move and bolstered his support. After becoming president, Caldera did just that — he dropped all charges against Chávez. Like most Venezuelan elites, Caldera viewed Chávez as a passing fad. Years afterwards, Caldera explained that nobody thought Chávez had even the remotest chance of becoming president.

But Caldera’s decision to free Chávez gave him credibility. On December 6, 1998, Chávez won the presidency by a large margin.

History of gatekeeping in the US

The US has an impressive record of gatekeeping, from both the Democrat and Republican parties. The US Founding Fathers actually designed the Constitution and electoral system for gatekeeping, as they did not fully trust the people’s ability to judge candidates’ fitness for office. They therefore came up with the Electoral College as a screening device.

The idea was that the Electoral College would be made up of locally prominent men in each state who would be responsible for choosing the president. But there were two problems:

  • The Electoral College only comes into play after the people vote.
  • In the early 1800s, the rise of political parties changed how the electoral system worked. The founders had originally envisaged that local notables would be the Electoral College’s delegates, but states began to elect party loyalists. This moved the gatekeeping function from the Electoral College to the parties.
    The parties have retained this gatekeeping function ever since.
1972 changes — binding presidential primaries

The turning point came in 1968, a year of riots and mass civil unrest following Martin Luther King Jr’s assassination in April and Robert F Kennedy’s assassination in June. The Democratic Party was deeply divided. The Democratic nominee, Hubert Humphrey, was popular with party insiders, but not with those opposed to the Vietnam War. He had also entered too late to run in a single primary, so his nomination significantly undermined the legitimacy of the party insider system.

After Humphrey lost the election, the McGovern–Fraser Commission reconsidered the nomination system and issued a set of recommendations that both parties adopted before the 1972 election.

From that point, presidential primaries became binding. Nominees no longer needed the support of the party establishment or gatekeepers. (In 1980, the Democratic Party backtracked somewhat by having “superdelegates”, which accounted for 15-20% of national delegates and which party leaders could use to fend off candidates they disapproved of. The GOP did not.)

Under the old system, between 1945 and 1968, only a single outsider (Dwight Eisenhower) publicly sought the nomination of either party. Whereas under the new primary system, between 1972 to 2016, 26 outsiders ran.

Money and media

Two further developments contributed to the continuing weakening of party gatekeepers:

  • The Supreme Court’s 2010 Citizens United ruling dramatically increased the availability of outside money;
  • The rise of alternative media, particularly social media and cable news (e.g. Fox News and influential radio talk-show personalities), tended to benefit ideologically extreme candidates.

It’s still not exactly easy for an outsider to win a nomination. Winning primaries still requires money, media coverage, and campaigners in all states. So the most successful outsiders are those who already have fame and/or money.

Trump vs the Republican establishment

Many people, including Nate Silver, believed that Trump’s chances of winning the Republican nomination were low because he did not have the support of party insiders. When the primary season began, Trump had no endorsements from Republican insiders.

But Trump had the early support of right-wing media personalities and attracted free mainstream coverage by creating controversy. He did surprisingly well in the state primaries, particularly on Super Tuesday (the day when the greatest number of states hold primaries).

The Republican establishment started freaking out — prominent Republicans like Mitt Romney, John McCain and Senator Lindsey Graham warned against Trump. But this was of little use. Under the primary-based system, Trump had the votes.

The authors argue that when Trump won the Republican nomination, the Republican establishment should’ve endorsed Hillary Clinton. While this may sound unthinkable, it has occurred in other countries before.

… when faced with a would-be authoritarian, establishment politicians must unambiguously reject him or her and do everything possible to defend democratic institutions—even if that means temporarily joining forces with bitter rivals
— Levitsky and Ziblatt in How Democracies Die

No prominent Republicans endorsed Clinton. Some refused to endorse Trump, but didn’t go so far as to endorse Clinton instead. Most prominent Republicans — such as Paul Ryan, Mitch McConnell, Marco Rubio, and Ted Cruz — did end up endorsing Trump.

If Republican leaders had the guts to endorse Clinton, the Republican electorate would have split. It would have sent a strong message that this was a moment of crisis, not an ordinary election. Even if just a small fraction of Republicans voted for Clinton, Trump would’ve lost.

Instead, the election became a standard two-party race. It was then essentially a toss-up between Trump and Clinton.

Unwritten democratic norms have been weakening

Two unwritten norms are fundamental to a functioning democracy:

  • Mutual toleration. This is the understanding that competing parties accept one another as legitimate rivals.
  • Institutional forbearance. This is the idea that politicians should exercise restraint in using their powers.
    The two are linked — if one party no longer accepts the other as a legitimate rival, they become more willing to do anything to win.

The authors argue that extreme partisan polarisation in the US has caused these two norms to weaken over time. This decline began when the two parties became divided not over mere policy, but over race and culture.

Constitutions are not enough

While constitutions usually have safeguards to prevent leaders from abusing power, they’re not enough by themselves. Constitutions are always incomplete — like any set of rules, there will be many gaps. And many constitutions confer virtually unchecked power during crises. [See further “How authoritarians consolidate power” below.] Germany’s 1919 Weimar constitution was designed by some of the country’s greatest legal minds, but quickly collapsed after Adolf Hitler gained power.

People can also interpret constitutions in different ways, following the letter of the law but not its spirit. For example, the US has few constitutional safeguards against filling agencies like the FBI with loyalists.

Democracies work best where constitutions are reinforced by strong democratic norms. Norms are informal rules that are not written down, but are still widely known and respected. They act as shared codes of conduct within a community or society. Norms can be hard to see, especially when they work well. But they play a key role in all parts of society, including politics. For example, much of the operations of the Senate and the Electoral College are governed by unwritten norms rather than formal rules.

Mutual toleration

Mutual toleration is the idea that as long as our rivals play by constitutional rules, we accept they have an equal right to exist, compete for power, and govern. Even if we disagree with or strongly dislike them, we accept them as legitimate rivals. Basically “agreeing to disagree”.

When the US was first founded, mutual toleration did not exist. Opposing those in power was considered treason. Both John Adams’ Federalists and Thomas Jefferson’s Republicans saw each other as a threat to the republic. Mutual toleration developed gradually, over many decades.

Institutional forbearance

Institutional forbearance is the idea that we avoid taking actions that comply with the letter of the law but violate its spirit. That is, not using dirty tricks or playing hardball.

Think of democracy as a game that we want to keep playing indefinitely. To ensure future rounds of the game, players must refrain from either incapacitating the other team or antagonizing them to such a degree, that they refuse to play again tomorrow.
— Levitsky and Zibaltt in How Democracies Die

The US system depends on a delicate balance between the executive branch, Congress, and the judiciary. Institutions that are strong enough to check the president need to regularly underuse that power.

Examples of institutional forbearance include:

  • Keeping to presidential term limits
  • Restrained use of Executive Orders
  • Not packing the Supreme Court
  • Confirming judicial appointments when they’re qualified
  • Using the filibuster sparingly
  • Impeaching a President only if there’s a reasonable chance he’ll be removed
Presidential term limits

For most of US history, the two-term limit was a norm, not a law. George Washington’s retirement after two terms set a powerful precedent.

Franklin Roosevelt broke this norm by serving four terms and is the only US president to have done so. But support for the two-term norm was very strong. Less than two years after Roosevelt’s death, a bipartisan coalition put it in the Constitution.

Executive Orders

Article II of the US Constitution lays out the formal powers of the presidency without clearly defining its limits. It is virtually silent on the president’s authority to act unilaterally via executive orders or decrees. Presidents can therefore circumvent the legislature by issuing executive orders, proclamations, directives, etc which do not require Congress’s approval.

George Washington worked hard to establish norms to stop the presidency becoming a new form of monarchy. He signed many bills he personally disagreed with and vetoed only two bills in his 8-year term (both times on constitutional grounds). In those 8 years, he issued only 8 executive orders.

During the 20th century, most US presidents were similarly restrained in using executive orders. Franklin Roosevelt broke this norm by issuing more than 3,000 executive orders — a record unmatched since. Many of these related to the New Deal and World War II. Still, Roosevelt’s acts triggered considerable bipartisan resistance.

Court packing

Authoritarians sometimes pack courts with judges who are loyal to them if they can’t remove judges directly. In Hungary, the Viktor Orbán government expanded the size of the Constitutional Court from 8 to 15. In Venezuela, the Hugo Chávez government expanded the Supreme Tribunal from 20 to 32.

There was some court-packing in the US in the early 19th century. But, by the late 19th century, it was widely viewed as unacceptable.

Example: Franklin Roosevelt’s attempt to pack the Supreme Court

In the 1930s, the Supreme Court declared several of Roosevelt’s New Deal programs unconstitutional. Roosevelt then made a bid to increase the number of Supreme Court Justices from 9 to up to 15.

Roosevelt was extremely popular and the Democrats had solid majorities. Yet his court-packing proposal was met with bipartisan opposition and fierce media criticism and ultimately failed. (It also helped that the Supreme Court itself reversed some of its previous decisions on New Deal legislation, making it easier for Democrats to oppose the court-packing plan.)

Confirming judicial appointments

While the President has the legal power to nominate justices, the Senate has the power of “advice and consent” over judicial appointments.

Historically, the Senate has exercised restraint in this power. Between 1880 and 1980, more than 90% of Supreme Court nominees were approved, and highly qualified nominees were always approved.

Example: Blocking Obama’s judicial nomination

After Justice Scalia died in 2016, President Barack Obama nominated appellate judge Merrick Garland to fill Scalia’s seat. Garland was undoubtedly highly qualified, and most viewed him as an ideological moderate.

For the first time in American history, the Senate refused to even consider his nomination on the grounds that it was too close to the next election. This was unprecedented. The Senate had never before stopped the president from filling a Supreme Court seat, no matter how close to the next election the vacancy arose.


The US Senate was intended to protect minorities from the power of majorities (which the founding fathers believed would be represented by the House of Representatives). They therefore designed the Senate to allow deliberation, with various tools (many unwritten) to slow down or block majority proposals. A prime example is the filibuster.

Without forbearance, the filibuster can easily lead to gridlock and dysfunction. For most of US history, this didn’t happen. There were only 23 obvious filibusters during the entire 19th century.

But filibuster use grew a lot starting in the 1980s, particularly in the Clinton (1993-2001), Bush (2001-2009) and Obama (2009-2017) presidencies. Between 2007 and 2012, there were 385 filibusters.


The US Constitution gives Congress the power to remove a sitting president via impeachment. Legally, impeachment only requires a simple majority in the House of Representatives, but conviction and removal of the impeached president requires a two-thirds vote in the Senate.

The norm is that the House of Representatives should not carry out an impeachment unless there is a reasonable chance it results in a removal.

Example: Richard Nixon bipartisan support for impeachment

After the Watergate scandal, nearly a dozen Republican senators joined Democrats in calling for an independent special prosecutor. When it was discovered that there were secret White House tapes that could implicate Nixon, leaders from both parties demanded that Nixon release them.

There was also more than enough Republican support to remove Nixon. (He was not ultimately impeached as he resigned first.)

Example: Bill Clinton impeachment

In 1998, the House of Representatives voted to impeach President Clinton for lying about his extramarital affair. The authors argue that this fell far short of conventional standards for impeachable offences, being high crimes and misdemeanours.

The Republicans also moved ahead with impeachment even though they didn’t have bipartisan support, such that the Senate almost certainly couldn’t convict him. Thus, for the first time in US history, the Republicans had politicised the impeachment process.

Mutual toleration and forbearance are linked

Mutual toleration and institutional forbearance are closely related and can reinforce each other. It’s easier to exercise forbearance when you see your opponent as legitimate. The converse is also true. If you see your opponent as a real threat, you may want to win at all costs.

In just about every case of democratic breakdown we have studied, would-be authoritarians have justified their consolidation of power by labeling their opponents as an existential threat.
— Levitsky and Ziblatt in How Democracies Die

This can lead to an escalating spiral.

The breakdown of democracy isn’t necessarily planned. It often begins slowly, with mere words. Each individual step seems minor. Some are taken under the guise of a legitimate aim — e.g. cracking down on corruption. Many citizens may not notice anything amiss initially as elections continue to be held. But it can soon escalate into a tit-for-tat between a norm-breaking leader and a threatened political establishment.

Example: Escalation in Peru

Peru’s Alberto Fujimori didn’t plan to be dictator, or even president. He started out as a little-known university rector, with no experience with legislative politics.

Fujimori believed the political elite as a corrupt oligarchy that was ruining the country. He was convinced that Peru needed drastic economic reforms and to increase counterterrorism measures, but had only a vague idea of how to do this.

Fujimori also lacked the patience for legislative politics. He preferred to govern alone, issuing decrees that the courts found to be unconstitutional. When his critics began accusing him of being “authoritarian”, he doubled down.

In 1992, Fujimori appeared on television and announced that he was dissolving congress and the constitution.

How did these two norms decline in the US?

There were several examples throughout the 20th century where democracy was challenged but effectively contained. We saw some examples above with Franklin Roosevelt and Richard Nixon.

Unfortunately, over the last 25 years, Democrat and Republican have become deeply divided. The authors believe the US’s extreme divisions are largely driven by race and religion. These deeply polarising issues tend to generate more tension than traditional policy issues like taxes and government spending.

Racial divisions

Paradoxically, the two democratic norms we’ve discussed emerged in the US out of profoundly undemocratic arrangements: racial exclusion and the consolidation of single-party rule in the South. When politics was restricted largely to whites, Democrats and Republicans had much in common. Neither party viewed the other as an existential threat.

Norms of mutual toleration started to form around the 1780s, but soon fell apart over the polarising issue of slavery. In the three decades before the American Civil War (1861-1865), there are estimated to be 125 violent incidents on the floor of the US House and Senate. We’re talking real violence here — politicians drawing pistols and even caning each other.

Democrats and Republicans gradually developed mutual toleration only after the issue of racial equality left the political agenda. This mutual toleration in turn encouraged forbearance. By the 1900s the two norms were well-established. The parties were far less polarised than they are today and frequently found common ground.

But the US wasn’t fully democratised until the civil rights movement, which culminated in the 1964 Civil Rights Act and 1965 Voting Rights Act. Full democratisation also led to polarisation. The Democrats became the party of civil rights while the Republicans became the party of racial status quo. Southern whites migrated to the Republican Party, while Southern blacks and northern liberals flocked to the Democrats.

Today, non-whites are disproportionately Democrat while the Republican Party has remained almost entirely white.


The Republican Party has also become the party of evangelical Christians. After Roe v. Wade legalised abortion in 1973, a large wave of evangelicals entered politics. The Republican party adopted increasingly pro-evangelical positions while the Democrats grew increasingly secular.


For most of the 20th century, both the Democratic and Republican parties were ideological “big tents,” each encompassing a range of political views.

By 2000, this changed. The Republicans had become mainly conservative while the Democrats were mainly liberal. Areas of overlap between the parties slowly disappeared, and they cooperated less frequently and voted consistently with their own party.

Polarisation since the 1970s

Polarisation grew after Newt Gingrich entered politics in the late-1970s. At the time, Republican leaders were friendly with Democrats and frequently cooperated with them. Gingrich thought they were “too soft”. He questioned Democrats’ patriotism and described them as “pathetic”, “sick” and “traitors”. Gingrich and his allies even distributed memos and audiotapes to spread these tactics across the Republican party.

When Gingrich became Speaker of the House in 1995, many younger Republican legislators saw him as a role model. The Republicans’ hardball approach became clear in Bill Clinton’s presidency, when they refused to compromise in budget negotiations, which led to government shutdowns in both 1995 and 1996. Clinton’s impeachment is another example [see above].

Things didn’t get better after Gingrich left Congress in 1999. President Bush focused on mobilising his base rather than independent voters, so he governed hard to the right. By this time, Democrats also favoured obstruction. They routinely filibustered Republican proposals and blocked an unprecedented number of Bush’s judicial nominees. Republicans reacted their own new tactics. The House introduced many more bills under “closed rules” which prohibited amendments, effectively abandoning oversight of the President.

The congressional watchdog became a lapdog, abdicating its institutional responsibilities
— Levitsky and Ziblatt in How Democracies Die

Things got worse during and after the 2008 presidential election between Obama and John McCain. Right-wing media outlets painted Obama as a Marxist, anti-American, and secretly Muslim. Prominent Republicans (e.g. Sarah Palin and Newt Gingrich) questioned his patriotism and legitimacy. Tea Party Republicans consistently claimed that he was not a “real American”. The birther movement went even further, questioning whether Obama was born in the US (with Donald Trump the most prominent birther). The claims were not confined to the fringes — a significant portion of Republican voters doubted Obama’s origins.

During President Obama’s term, Republicans embraced obstructionism, especially after many Tea Party Republicans entered Congress in 2010:

  • Filibuster use skyrocketed
  • The confirmation rate for judicial appointments plummeted.
  • In 2011, Republicans refused to raise the federal debt ceiling until Democrats agreed to certain demands. This was unprecedent — raising the debt limit was a longstanding bipartisan practice.
  • In 2015, Mitch McConnell urged US states to ignore Obama’s regulatory order limiting greenhouse gas emissions.
  • Nearly 50 Republican senators wrote an open letter to Iran’s leaders saying that Obama had no authority to negotiate a deal over their nuclear program. Though diplomatic negotiations have long been the responsibility of the executive, the Republicans were angered by Obama’s decision to use an “executive agreement” instead of a treaty.

The Democrats responded with their own norm-breaking. In 2013, Senate Democrats voted to eliminate the filibuster for most presidential nominations, a move that many called the “nuclear option”. And Obama himself issued plenty of unilateral executive actions.


Trump’s norm-breaking is unprecedented:

  • After he won the 2016 election, Trump continued to attack both Hillary Clinton and Obama, undermining mutual toleration and norms of post-election reconciliation.
  • Trump broke longstanding norms around separating private and public affairs, nepotism and conflicts of interests by refusing to divest his businesses.
  • Trump was the first major politician in over a century to challenge the integrity of the electoral process.
  • The authors believe Trump’s brazen lying is one of the most notorious forms of norm-breaking. Lying is corrosive to a political system as it erodes trust in government and jeopardises our access to credible information. While politicians have never been known for being truthful, they usually avoid direct lies like Trump’s. During the 2016 campaign, PolitiFact classified 69% of his public statements as “mostly false” or worse, while only 17% were “mostly true” or better.
  • Trump publicly insulted media outlets and even individual journalists. His administration also broke established norms by selectively excluding reporters from press events and handpicking journalists from smaller but sympathetic outlets.

All of this is certain to weaken norms in the long run. Humans have a limited ability to cope with repeated norm-breaking. When it happens over and over, we become desensitised and respond by lowering our standards of what counts as “deviancy”.

Norms are the soft guardrails of democracy. As they break down, the risk of the hard guardrails breaking down increases. So even if Trump does not break the hard guardrails of democracy, he’s increased the risk that a future president will.

Why did the Republicans break more norms?

While the Democrats broke some norms, overall the Republicans showed less mutual toleration or institutional forbearance.

The authors believe this is because:

  • Media outlets. Highly partisan media outlets play a much bigger role on the right. [There’s some evidence to suggest that mainstream media has a slight left bias in terms of words used, albeit not in the stories they choose to cover. I wonder if this slight left bias encourages the development of more partisan-right outlets?]
  • Well-funded interest groups. Conservative interest groups push the Republican party further to the right. When campaign finance laws loosened after the 2010 Citizens United decision, these groups gained greater influence in the Republican party.
  • White Protestant majority declining. For nearly 200 years, white Protestants were the majority in the US and dominated politically, economically and culturally. Today, they are a shrinking minority — and have found refuge in the Republican Party. The authors believe this is a large part of why there’s such intense animosity in the American Right. It may also help explain the discourse around who is a “real American” and the appeal of slogans such as “Make America Great Again”.

How authoritarians consolidate power

Earlier, we looked at four authoritarian warnings signs that helped predict whether a politician will become an authoritarian once in power. After coming into power, an authoritarian may take various actions to consolidate power:

  • Capturing the referees and judges;
  • Sidelining the opposition;
  • Tilting the playing field; and
  • Exploiting a crisis.
… would-be authoritarians must capture the referees, sideline at least some of the other side’s star players, and rewrite the rules of the game to lock in their advantage, in effect tilting the playing field against their opponents.
— Levitsky and Ziblatt in How Democracies Die

Capture the referees

Democracies have certain institutions that are meant to be neutral: the courts, law enforcement, and the intelligence, tax, and regulatory agencies. These institutions generally have the authority to investigate or punish wrongdoing, whether it’s wrongdoing by public officials or private citizens. [As a non-American, I find it weird how the appointment of judges in the US is so politicised. It undermines the separation of powers, yet it’s built directly into the US Constitution.]

Capturing the referees usually involves quietly firing civil servants and replacing them with loyalists. They can then use these referees either:

  • as a shield (so they themselves are not subject to the law), or
  • as a weapon (to selectively enforce the law against their opponents).

For example, President Putin set the tax authorities after Vladimir Gusinsky, after Gusinsky’s independent TV network became a “pain in the neck” for Putin.

Example: Capturing the judges in Argentina

When Juan Perón became Argentina’s president, 4 of its 5 Supreme Court justices were conservative opponents.

Perón and his allies then impeached 3 of them on the grounds of malfeasance (the fourth resigned before he could be impeached), and appointed 4 loyalists in their place. The court never opposed Perón again.

Sideline the opposition

In the past, dictators like Benito Mussolini or Fidel Castro jailed or even killed their rivals. Today, autocrats tend to hide behind a veneer of legality. They might not wipe out all traces of dissent, but will ensure that key players — opposition politicians, business leaders, major media outlets, and influential public figures — are sidelined or neutered.

One way they do this is by buying them off with public positions, favours or outright bribes. Friendly media outlets may get privileged access to the president. Friendly business leaders may receive government contracts.

A common way to sideline the opposition media is through libel or defamation suits (which is particularly effective if you’ve captured the referees). This can have a chilling effect — media outlets start to self-censor to avoid the risk of being sued.

Example: Sidelining the business leaders in Russia

President Putin summoned 21 of Russia’s wealthiest businessmen and told them they’d be free to make money under his regime, so long as they stayed out of politics. Most heeded his warning. Mikhail Khodorkovsky, head of a giant oil company and possibly Russia’s wealthiest man at the time, did not. Khodorkovsky was a liberal who disliked Putin, and began to finance opposition parties.

Putin had Khodorkovsky arrested in 2003 for tax evasion, embezzlement, and fraud. He was imprisoned for nearly a decade. [In 2003, Forbes estimated Khodorkovsky was worth around $15 billion. At the end of 2013, after he left Russia, he was estimated to be worth only $100 to $250 million. He’s now living in exile in London.]

Tilt the playing field

Authoritarians can tilt the playing field in their favour by changing the constitution, the electoral system, and other institutions in ways that lock-in their advantage for years or even decades.

Often the changes are under the guise of some public good. One of the most striking examples comes from the US.

Example: Black disenfranchisement

Many African Americans became eligible to vote after the Civil War, and they overwhelmingly voted Republican [note: the Republican and Democratic parties flipped ideologies around the 1960s and 1970s]. This posed a major threat to southern white political control and to the Democratic Party, especially as African Americans made up a majority or near-majority in many southern states.

So all 11 post-Confederate states reformed their constitutions and electoral laws to disenfranchise African Americans. They did so by introducing seemingly “neutral” poll taxes, property requirements, literacy tests, and complex written ballots.

Black voter turnout, which had reached 96% in 1876, plummeted to 11% by 1898.

Exploit a crisis

Most constitutions expand executive power during crisis, so autocrats often use these to justify antidemocratic measures. Judges are also highly reluctant to block presidential power grabs when national security seems to be at risk. The public is more likely to tolerate or even support authoritarian measures when they fear for their own safety. For example, after the terrorist attacks on September 11, 2001, President Bush’s approval rating skyrocketed to 90%.

Crises may include an economic crisis, natural disaster, and especially security threats (e.g. war, terrorism, etc). Some leaders even go so far as to invent a crisis to cling onto power — a famous example is the Reichstag Fire, which happened just a month after Hitler was sworn in as chancellor.

What is the solution?

There is a growing sense that democracy is in retreat all over the world, with some saying we have entered a period of “democratic recession”. The authors were generally sceptical of this idea, at least before 2016. While democracies have declined in some countries, there are others (e.g. Colombia, Sri Lanka and Tunisia) that have grown more democratic.

But in writing this book, the authors have realised that American democracy is not as exceptional or resilient as we may think. The authors are also worried about Trump’s rise and what that means for global democracy. Since the Cold War, the US has generally promoted democracy as part of its foreign policy but the US under Trump seems to be abandoning this role. Moreover, America itself is no longer a democratic model.

The authors see little hope of a swift recovery for democracy, even if Trump fails politically. The soft guardrails of democracy have been weakening for decades, and removing Trump will not suddenly restore them. In their view, the most likely future after Trump future is one with increased polarisation and departures from unwritten norms — democracy without solid guardrails.

Democracy is a shared enterprise. Its fate depends on all of us.
— Levitsky and Ziblatt in How Democracies Die

So what can we do?

Don’t fight dirty

After the 2016 election, some argued that Democrats needed to “fight like Republicans” and do whatever they can prevent judicial appointments and block laws, regardless of their merits.

The authors disagree with this approach. Fighting dirty just scares off moderates and gives the other side greater justification for cracking down. It often plays directly into an authoritarian government’s hands. Moreover, even if Democrats won using hardball tactics, they would inherit a democracy without any guardrails so the victory would be Pyrrhic. In the longer run, it could lead to a president even more dangerous than Trump.

Example: Scorched-earth tactics in Venezuela

Opposition to the Chávez government in Venezuela tried several things:

  • backing a military coup;
  • launching an indefinite general strike, seeking to shut the country down until Chávez resigned; and
  • boycotting the 2005 legislative elections.

All three of these backfired. The coup failed and destroyed the opposition’s image as democrats. The strike lasted 2 months, costing Venezuela an estimated $4.5 billion but still ultimately failing. Moreover, the tactics eroded the opposition’s public support and gave the government an excuse to purge the military, the police, and the courts, arrest or exile dissidents, and close independent media outlets.

The authors argue that any opposition should seek to preserve and strengthen democratic rules and norms. That means focusing on legal, democratic institutions like Congress, the courts, and elections, and engaging in peaceful, not violent, protests.

Don’t eschew diversity

As noted above, political divisions increased and norms of mutual toleration started breaking down after the two parties became divided by race. It may therefore be tempting to think that the Democrats should reduce the influence of ethnic minorities and become less accepting of both immigration and racial equality to win back the white working class.

The authors believe this is a terrible idea. It might reduce polarisation, but would repeat some of the US’s most shameful mistakes in doing so.

Build a broad pro-democratic coalition

The most effective coalitions are those that bring together groups that don’t agree on many issues, but do agree on one important one. A broad anti-Trump coalition would likely require progressives teaming up with business executives, religious leaders (particularly white evangelical ones), and red-state Republicans.

A broad coalition that cut across the current axes of partisan division could perhaps help reduce them. Mutual toleration thrives when we line up on different sides of issues with different people at different times. If we agree with our political rivals at least some of the time, we’re less likely to view them as enemies.

In contrast, if the anti-Trump coalition only accepted those who held the same views on abortion rights or healthcare, it would be much narrower. If it only had urban, secular and progressive people, it would reinforce, rather than break down, America’s existing divisions.

[T]he fundamental problem facing American democracy remains extreme partisan division—one fueled not just by policy differences but by deeper sources of resentment, including racial and religious differences.
—Levitsky and Ziblatt in How Democracies Die

Reform the Republican Party

The main driver of the chasm between the parties today is largely due to the Republican Party. Over the last 25 years, the party has moved increasingly towards the right. The party has become dysfunctional as its core and leadership structure have been “eviscerated”. Outsiders (either wealthy donors, well-funded interest groups, or right-wing media) now hold much greater sway over the party than the Republican Party’s own leaders. All of this has made the Republican Party vulnerable to takeover by extremists.

The Republican Party must therefore be reformed. Leaders must regain control in the four key areas of finance, grassroots organisation, messaging and candidate selection. They must broaden their base beyond white Christians. Republicans also need to expel extremists from their ranks and distance themselves from Trump.

Focus on depolarising economic policies

The authors recommend focusing on addressing economic inequality in a way that is depolarising. For example, means testing (giving benefits only to those who fall below an income or asset threshold) is often polarising, as it makes many middle-class voters believe that only poor people benefit from social policy. Such policies can also be racially stigmatising, as race and poverty have historically overlapped in the US.

Policies that follow the more universal models in northern Europe may be less polarising. Examples include comprehensive health insurance or universal basic income (UBI). (The Nixon administration once seriously considered a UBI and even introduced it into Congress.) [This was the Family Assistance Plan, a negative income tax, which is not really the same as a UBI.]

While the authors admit they can’t be sure that these universalistic policies would contribute to a broad, durable coalition, they think they stand a better chance than the current means-tested programs.

Other Interesting Points

  • Presidential systems are more susceptible to populism than parliamentary democracies. In parliamentary democracies, the prime minister is selected by the leading parties in parliament so they will be acceptable to political insiders. In contrast, Congress does not elect the president. At least in theory, presidents are elected by the people, and anyone can run for president.
  • Before 1840, campaigning had not been the norm in US elections. Candidates were supposed to act like they had no personal ambition for power. William Henry Harrison broke this norm by actively campaigning for the 1840 presidential election.

My Review of How Democracies Die

How Democracies Die focuses on US politics, particularly Trump, so is inevitably going to be divisive. I’m sure it will have far more readers on the left than on the right.

As a non-American, I found the book to be educational. There are many features of America’s political system that I find very strange, like the primary system. Or political appointments to the Supreme Court. Or the fact that people register as Republicans or Democrats (and openly describe themselves as such). I really don’t know what to make of US politics. Most of what I’ve read makes it look like such a mess that I’m inclined to believe it’s slightly less dysfunctional than the media often portrays. But then this book does make things look pretty dire.

It’s hard for me to assess how balanced the book was. Some of the language is certainly more emotive than it strictly needs to be (“stealing” a Supreme Court seat, “trampled” a norm, holding “hostage”) and overall there is definitely more blame meted out to Republicans. Yet the authors back up their claims with facts (some of which I checked) so this may well be warranted. However, I am far from an expert on American politics or history, so wouldn’t know if examples were cherry-picked. So — I think the book was balanced, but don’t know enough to really say.

For me, the two most surprising parts of the book were:

  • Gatekeeping. The authors seem to fundamentally believe that party leaders should play a strong gatekeeping role, even as they admit that gatekeeping is undemocratic. I’m not sure how I feel about that. I’m certainly not a democratic purist but gatekeeping seems to involve a real risk of cronyism and subverting the will of the people. That said, parliamentary democracies seem to involve even more gatekeeping and I’ve never found that problematic, so my own biases may be creeping in here.
  • The relationship between black enfranchisement and the US’s political divisions. I’m probably naïve about race relations in America but I hadn’t really thought about race being such a big driver of the US’s political polarisation. It certainly seems plausible. I think many well-intentioned people often downplay or overlook racial differences, but that can also be dangerous. In Political Tribes for example, Amy Chua argues that America’s failure to understand the existing tribal divisions in other countries led it to make various foreign policy mistakes.

One problem with How Democracies Die is that the solutions it offers are rather weak. It’s not really the authors’ fault, because the deep divisions in America are incredibly hard to solve. I liked, and strongly agree with, their suggestion to build a broad pro-democratic coalition. But their argument that universalistic policies like a UBI would be less polarising than means-testing seemed particularly ill thought-out. I generally like the idea of universal policies myself, but the problem is always paying for them. A UBI requires much higher taxes than means-testing, and taxes tend to be very polarising. Yet the authors don’t even mention this at all. Like many books about wicked problems (e.g. Evicted), How Democracies Die is a bit better at diagnosis than cure.

What did you think of this How Democracies Die? Did you find it balanced? Share your thoughts in the comments below!

Buy How Democracies Die at: Amazon | Kobo (affiliate links) <– These are affiliate links, which means I may earn a small commission if you make a purchase. Thanks for supporting the site! 🙂

If you enjoyed this summary, you may also like:

6 thoughts on “Book Summary: How Democracies Die by Steven Levitsky and Daniel Ziblatt

  1. Thanks, this is really interesting.

    Minor thing: I’ve only learned this recently but abortion was not an evangelical issue until about 5 years after Roe v Wade. (The story of why it became an issue is actually crazy – see link below). Until that time, it was a Roman Catholic issue and one that evangelicals just didn’t really care about it. The general evangelical position on abortion then was what is an (outside the US today) a moderate conservative position that there should be some limits but of course it is justifiable in quite a few circumstances.

    This story is told in a few places. One is the very short book *Bad Faith* which I have not read. But it’s also told in the first episode of the excellent BBC podcast, *Things Fell Apart*, which is about the strange origins of current culture war issues.

    The whole podcast is great. The presenter of that podcast, Jon Ronson, is also the author of an excellent book called *So You’ve Been Publicly Shamed” about social media shaming which I read a few years ago and really enjoyed.

    1. Thanks for the podcast recommendation – I listened to it today. It’s a great example of path dependency in the moral domain.

      Also interesting how abortion became an issue that galvanised evangelical Christians into the organised political bloc they are today. Catholics by contrast are a much broader population with diverse range of views, and individuals are likely to line up with different political parties on different issues depending on what they prioritise. So it’s harder to unify them into a political bloc.

      1. Yes, the path dependence is crazy. Seems that the key reason it started to become an evangelical issue was the existing evangelical disdain for feminist causes (women putting careers first etc.) So once the feminist groups aligned themselves as pro-choice in a confrontational way, the evangelicals “naturally” took the other, anti-abortion side.

        Reminds me of the surprising ideological reversals during the early days of covid. As far as I remember, originally shutting the borders was more of a “right-wing” idea. Because it sounds anti-immigration, various left-wing groups thought it was bad (e.g. this article on stuff from March 2020:–why-we-have-done-so-much)

        But then once the left-wing coded public health people said it was a good idea, the right-wing people flipped and began to say that covid wasn’t such a big deal. I think for the same reasons that the evangelicals took up abortion – they didn’t like the people for other reasons.

        The same thing happened with masks – the science people said at first there was no evidence they worked, and wearing masks seemed to be a right-wing “dumb anti-sicence” thing to do, but then it flipped and it became a left-wing thing to do and the right-wingers hated it.

        In your review you mention the flipping of ideological views in the 60s and 70s between republicans and democrats, I wonder if this dynamic – choosing to take the other side of an issue if your existing political enemies have a view – might explain it as different issues cycle through political importance.

        1. Yeah, I recently read The Righteous Mind by Jonathan Haidt who talks about how we like to think our moral positions are based on reason and logic but for most people it’s not really true. He uses the analogy of an elephant (emotions) and its rider (logic) — the rider has some sway, but the elephant really chooses where to go.

          He then explains that when we change our minds on moral issues, it’s usually not because of the rider (i.e. because we are persuaded by the rationality of a brilliant argument) but because of friendly elephants (i.e. we find out someone we like holds a view we disagree with and we naturally start being more open to that view). It’s why I really liked Levitsky and Ziblatt’s recommendation to build a broad pro-democratic coalition.

          The opposite seems to apply to “hostile elephants” (though Haidt doesn’t discuss this). The whole “hating abortion because feminists support it” is a great example of it, and Covid too as you point out. Like, most of us simply don’t have well-informed views on epidemiology or the science behind vaccines and masks. Instead, we form our views based on who we trust — and right- and left- people trust different people and different institutions.

  2. I remember talking with a Burmese friend of mine when the Myanmar coup first began a few years back, and to my naive USA-centric self, it was eye-opening to see how “real” authoritative threats can be, how delicate democracy is. I think it’s something those who were born and raised in Western societies take for granted. Just think of how much time has elapsed since the Cold War – this past generation knows of nothing but relative stability and peace.
    Part of me wonders if this is cyclical, sort of like how World War I was spurred by a desire to flex muscles & fight by a generation of people who never experienced the terrors of war (I’m sure this is reductionist, but I hope the general point conveys).

    1. Good point about the cyclical nature of history. Chris Blattman has also talked about how young hotheads in Colombia are likely to agitate for war but are stopped by older gang members who know how costly war is so try to avert it. You see similar things happen in the economy – those who have lived through stockmarket bubbles and crashes are more reluctant to invest in stocks and more likely to prefer housing, which may contribute to a housing bubble/crash. (That said, there are many causes of high house prices so I don’t want to be reductionist here, either!)

      Our individual lifespans are short and we learn more effectively from our own experiences than those of others. So I believe we are doomed to repeat at least some of our past mistakes. But as a species we do manage to learn collectively, albeit imperfectly, so taking a broader view I’m optimistic there is generally an upward trajectory though we might slip along the way. (That said, those slips can ruin lives and take decades to claw back, so I don’t want my general optimism to suggest we should be complacent, either.)

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.