The romantic view of coalition collapse involves a dramatic walkout, a slammed door, a junior partner storming from the cabinet table in principled fury. The reality is considerably less cinematic. Coalition governments die by a thousand cuts — through accumulated slights, ministerial turf wars, and the slow poisoning of relationships between people who never particularly liked each other but agreed to share power anyway.
This matters because coalition government is now the norm, not the exception. In the European Union, single-party majority governments have become rare specimens. Germany, Italy, the Netherlands, Belgium, Israel, and dozens of other democracies routinely produce parliaments where no party commands a majority, forcing erstwhile rivals into awkward political marriages. Understanding how these arrangements fail is essential to understanding modern democracy itself.
The three-body problem
Coalitions of two parties are inherently more stable than coalitions of three or more, and mathematics explains why. With two parties, there is one relationship to manage. With three parties, there are three bilateral relationships. With four, there are six. The complexity grows exponentially while the political bandwidth to manage it remains fixed.
This is why the most spectacular coalition failures tend to occur in fragmented parliaments. Israel's revolving-door governments stem partly from an electoral threshold so low that a dozen parties routinely win seats. The Netherlands has spent months at a time in caretaker status while negotiators attempt to assemble viable combinations from parliamentary confetti. Belgium once went well over a year without a formal government — a record that would have seemed absurd had anyone predicted it.
The junior partner's dilemma
Smaller coalition partners face an existential trap. They must demonstrate relevance to their voters by extracting policy concessions and visible wins. But every concession they extract comes at the expense of the senior partner, who commands more seats precisely because more voters preferred their platform. The junior partner's success breeds the senior partner's resentment.
Worse, junior partners who govern competently tend to see their vote share absorbed by the larger ally. Voters conclude that if the coalition works, they might as well vote for the party actually running things. Junior partners who govern obstructively preserve their identity but destroy the government's functionality. There is no comfortable middle ground, which is why small parties so often either disappear after coalition participation or blow up the arrangement to preserve their distinctiveness.
The confidence vote as theatre
Formal confidence votes rarely cause coalition deaths; they merely certify them. By the time a vote is called, the outcome is typically known. The real action happens in the weeks before, in private negotiations over whether the coalition can be salvaged or whether partners have already begun positioning for the election to follow.
Smart coalition architects build in pressure-release mechanisms — regular summits between party leaders, agreed procedures for resolving disputes, even scheduled renegotiations of the coalition agreement. The German model of detailed written coalition contracts, sometimes running to hundreds of pages, represents one approach. The Italian model of creative ambiguity represents another. Neither guarantees survival, but both acknowledge that coalitions require active maintenance rather than mere initial agreement.
Our take
The fragility of coalition government is a feature, not a bug. Systems that make power-sharing difficult also make power concentration difficult. The price of preventing any single faction from dominating is accepting that governance will sometimes be slow, frustrating, and prone to sudden collapse. Voters in coalition-dependent democracies have implicitly accepted this bargain, even if they complain about the results. The alternative — winner-take-all systems that produce decisive governments — carries its own costs, including the periodic entrenchment of parties that half the country actively despises. There is no free lunch in constitutional design, only different ways of distributing the bill.
