Why do shifts in public opinion often not translate into policy?  Why and when can parties in (and out of) government move policy toward their preferences?  In parliamentary democracies, which are the majority of democracies, coalition governments are common and outcomes are the product of bargaining between parties. Scholars have proposed several means of predicting which parties are most likely to prevail in bargaining, ranging from polls to bargaining power indices. Few of these measures, however, have much predictive power and most — the exception is polls — do not offer values between elections, falsely suggesting that coalition bargaining does not continue over the life of a government. We develop a dynamic empirical measure of parties’ bargaining leverage that can predict which party’s preferences make it into policy and laws. It should be useful for researchers interested in policy-making, pledge-keeping, coalition politics, representation and other issues. We argue and demonstrate that parties’ probability of government inclusion, were a new government to be formed, determines the credibility and efficacy of threats to leave the current government.  Parties that can credibly threaten to abandon a government and form another, especially one excluding their current bargaining partner, have considerable bargaining leverage.