In December 2016, Ryszard Petru, center, leader of the opposition party Nowoczesna, holds a sign advocating press freedom along with other lawmakers during a protest in the Sejm, the lower house of the Polish Parliament in Warsaw. The opposition’s occupation of the Parliament building ended Jan. 11. (Marcin Obara/EPA)

Democracy is under attack in Poland, and in other countries as well. For several months, Poland’s ruling Law and Justice (PiS) party has pushed for new limits on judicial independence that would greatly expand the party’s power.

Poles pushed back with mass street protests in July, and the European Union threatened to suspend Poland’s E.U. voting rights — and Polish President Andrzej Duda vetoed key portions of the reforms. In September, Duda floated a compromise plan that would force several prominent judges to retire but avoids giving PiS complete control over their replacements.

The battle over Poland’s judiciary is one chapter in the widespread struggles over the concentration of power in the hands of executives and ruling parties in countries such as Hungary, Turkey, Nicaragua, Venezuela and Russia. Increasingly, democracies erode not from military intervention or revolutions, but from the expansion and abuse of power by elected leaders. Executive power remains an ongoing political fight in the United States and other established democracies as well.

Power-sharing preserves democracy

In a recently published paper, we show why reining in central government power is so critical to democratic survival. A core question that has troubled democratic designers from the very beginning is how to empower election winners — but still protect election losers and minority groups.

When unconstrained, elected leaders are often tempted to consolidate their power by shutting out minorities and other opposition groups. In Burma, Iraq and Thailand, for instance,  democracy has struggled to take root when alienated populations are frozen out from electoral power.

What “power-sharing” strategies work best to solve this democratic design problem by sharing, distributing or limiting political power?

We compare three distinct approaches: (1) Inclusive power-sharing that guarantees groups a share of central government power, such as through ethnic quotas; (2) Dispersive power-sharing that divides power into separate spheres, such as through federalism; and (3) Constraining power-sharing that limits government power and gives greater autonomy to individuals and groups, such as through independent judiciaries and religious protections.

We looked at the last 40 years to compare how successful each strategy proved to be at securing democracy. Most of the time, inclusive and dispersive power-sharing don’t seem to matter much for democratic survival. An exception is in countries recovering from civil war, where inclusive power-sharing helps democratic survival and dispersive power-sharing harms it.

Political constraints prevent abuse of power

Our most significant finding is that political constraints are the strongest and most consistent predictors of democratic survival. As seen in Ukraine, Nigeria and the Philippines, strong, independent courts can be highly effective at preventing executives from abusing their power, overturning term limits and even stealing elections.

In turn, if election losers know that their rights are protected, they are less likely to take up arms and challenge election results by force. Instead, they can organize peacefully to contest the next election, with greater confidence that the electoral process will run fairly.

Countries such as Liberia, with a recent history of civil war, see the positive effects of constraining power-sharing particularly strongly. In these countries, most political groups already have experience on the battlefield, making it easy for them to return to violence if they can’t achieve their aims at the ballot box. Yet Liberian President Ellen Johnson Sirleaf confidently told the United Nations that last week’s elections place Liberia on an “irreversible course” toward democratic consolidation.

Liberia has been a surprising democratic success story, in part due to its relatively strong executive constraints. For example, the Supreme Court has been active in reviewing decisions by the National Elections Commission to disqualify several prominent candidates. In two cases, the court overturned commission decisions, ruling that previously disqualified candidates must be allowed to run. This sends a strong signal about the best strategies for seeking power. If opposition candidates are free to contest elections, they are less likely to grab power through renewed civil war.

How does this analysis apply to Poland?

Our results on government constraints suggest just why Law and Justice’s attempted control over the judiciary (as well as limits on the media) is so dangerous for democracy in Poland. It also explains why this control is in the party’s interest: Checks and balances prevent ambitious governing parties from rapidly reshaping a country’s politics. The strong E.U. reaction, including a threat to send the case to the European Court of Justice, reflects the significance of the reforms.

The findings also can apply more broadly to democratic design. Domestic and international actors deciding on post-conflict constitutions or invested in political reforms can place renewed attention on the role of political constraints. In particular, our research suggests that shifting away from dispersive power-sharing (especially federalism) and toward strong constraints is probably a more effective strategy for securing democracy. Inclusive arrangements may support democracy following severe conflicts but are often counterproductive over the long term.

Many political thinkers over the years have seen checks and balances as central to democratic consolidation. Yet in new democracies, establishing effective constraints and rule of law often takes a back seat to electoral administration and intergroup bargaining.

We find that limiting executive power both lowers the stakes of elections and protects vulnerable minority groups from abuses of power. For more established democracies, these constraints aim at the greatest contemporary threat to democracy: abuse and expansion of power by elected officials.

Above all, successful democracy is not solely about conducting elections, but also requires that the consequences of elections are tolerable to all sides, protect individual rights and ensure fair party competition in the future.

Benjamin A. T. Graham is an assistant professor in the School of International Relations at the University of Southern California and co-founder of the Security and Political Economy (SPEC) Lab.

 Michael K. Miller is an associate professor of political science at George Washington University, studying democratization, autocratic institutions and democratic survival.

Kaare Strøm is the distinguished professor of political science at the University of California at San Diego.