In many parliamentary democracies, only strong institutions can solve the coalition dilemma and help coalition partners control ministerial autonomy and implement a coalition compromise. We develop a new theory of parliamentary scrutiny in coalition governance with state-dependent preferences that shows under which conditions coalition compromises remain enforceable during a government term. Relying on the informational differences within coalition governments, our theory considers the risk of inaction when coalition conflict reaches a critical point or when a coalition party does not accept ministerial autonomy due to low inaction costs. We test the empirical implications of our new theory using the dataset on coalition governance by Martin and Vanberg (2011). Our findings will show whether large partners are more effective in controlling ministerial drift or promote coalition inaction when a minister from a small party expects being scrutinized. We will also investigate the type of coalition compromise, which we expect being more vague when coalition conflict increases.
