The Cost of Ruling

cost of ruling (n.): The empirically observed tendency of governing parties to lose support simply by governing. Not to be confused with governing badly, which also loses support but is a separate, more voluntary achievement.

Political scientists have been documenting this pattern since at least the 1980s. In 2003, Peter Nannestad and Martin Paldam analyzed 282 elections across 19 democracies¹ and found that governing parties lose approximately 2.25 percentage points of vote share per election cycle. They called this “an unusually stable constant.” In 2022, Alfred Cuzán confirmed it with a larger dataset: two to three points in OECD democracies, like clockwork.² Multiple scholars now call this “one of the few empirical laws of political science.”³ The loss is close to or larger than the plurality margin in most recent elections, meaning it’s often enough to flip control of a government.

Three mechanisms do the grinding.

Over-promising. Parties say whatever it takes to win. Governing requires choosing, and choosing means disappointing somebody. The opposition gets to stand at the edge of the pool looking dry and appealing, offering what Nannestad and Paldam described as alternatives that “shine in comparison to the compromises necessary when they later must balance the state budget.”

Policy drift. Once in office, parties move policy toward their own preferences, because that’s what parties do. The median voter quietly wanders toward whoever isn’t doing that. Christopher Wlezien found that this mechanism accounts for roughly half the total cost of ruling in U.S. presidential elections.⁴ Matt Grossmann and Wlezien later estimated that each net ideological law costs the governing party about a quarter of a percentage point of congressional vote share.⁵ Legislate successfully, and you are rewarded with losing support. Legislate unsuccessfully, and you get the same prize.

Grievance accumulation. Human brains store bad news more readily than good news. A cut heals; the scar reminds you. Even in a growing economy with falling unemployment, the running tally of disappointments only grows. The media amplifies this: in a study spanning four European countries over two decades, Gunnar Thesen, Christoffer Green-Pedersen, and Peter Mortensen found that the extra media coverage incumbents receive actually skews negative — a burden, not a bonus. Accounting for this media effect cut the estimated impact of time-in-office by half.³

The pattern is remarkably indifferent to institutional design. In Cuzán’s expanded dataset, parliamentary systems pay about two points per cycle; presidential systems closer to three. It holds at the subnational level too: U.S. gubernatorial races show a cost of about 3.2 points across nearly a thousand elections. Outside the OECD, the cost runs much steeper: about five points in newer parliamentary democracies, and nearly nine in Latin American presidential systems, as though democracy charges a higher fare in more turbulent economies.²

One variable moderates it: how many parties share power. Single-party governments pay the full price. Coalition governments split the bill, because voters can punish one partner without tossing the whole coalition.⁴

The cost of ruling is what keeps democracies from becoming dictatorships and dictators from becoming democrats. It is the bill that arrives for winning. And it usually comes due the next election, regardless of the services provided.