The last decade has witnessed a proliferation of scholarly works on the effects of divided government on government performance in Taiwan. Vast empirical evidence asserts that it has a negative impact. This assertion is puzzling from the perspective of the pivotal politics model, as the Legislative Yuan seems to be the only pivot in the Taiwanese law-making process and, therefore, should be able to dictate any policy changes. This paper attempts to resolve this discrepancy by developing four game-theoretic models to examine under what conditions legislative gridlock can arise. Innovatively, I go beyond previous spatial models by incorporating position-taking and legislative inactive incentives into traditional pivotal politics models. Our key finding is that legislative stalemate can arise only when the majority party in the legislature becomes inactive. One of the incentives for legislative inactiveness is Sundquist's (1988) presumption about the antagonistic legislative-executive relationship (i.e., each branch intends to defeat each other) under divided government. Nonetheless, the negative effects of divided government can be minimal or moderate if the executive can counteract legislative inactiveness by successfully going public. Finally, we discuss the empirical implications of the theoretical models for passage of government initiatives and possible solutions to legislative gridlock.