Abstract
We develop a game-theoretic model of strategic interdependence and tipping in public policy choices and show that the model can be estimated by probit and logit estimators. We test its validity and applicability by using daily data on state-level COVID-19 responses in the United States. Social distancing via shelter-in-place (SIP) strategies and wearing masks emerged as the most effective nonpharmaceutical ways of combatting COVID-19. In the United States, choices about these policies are made by individual states. We develop a game-theoretic model of such choices and test it econometrically, confirming strong interdependence in the implementation of these policies. If enough states engage in social distancing or mask wearing, others will be tipped to follow suit. Policy choices are influenced mainly by the choices of other states, especially those of similar political orientation and to a lesser degree by the number of new COVID-19 cases. The choice of mask-wearing policies is more sensitive to peer choices than the choice of SIP policies, and Republican states are much less likely than Democratic to introduce mask-wearing policies. The choices of policies are influenced more by political than public health considerations. These findings emphasize strategic interdependence in policy choice and offer an analytical framework for these complex dynamics.