If different producer groups have divergent interests concerning macroeconomic policies, how do societal preferences translate into state policy outcomes? I develop and test a party-as-agent framework for understanding the importance of societal preferences with regard to monetary policy under capital mobility. Following the principal-agent model, political parties function as agents for different societal principals. Rightist parties tend to represent internationally oriented business groups with preferences for monetary convergence, while leftist parties do the same for domestically oriented groups preferring monetary autonomy under capital mobility. I present statistical evidence showing that OECD leftist governments have been associated with more monetary autonomy and currency variability than their rightist counterparts, even after controlling for basic economic indicators such as inflation. The statistical evidence also shows that societal group size tends not to explain either autonomous monetary policy choices or exchangerate stability. Thus even large and wealthy societal groups may be unable to obtain their preferred policy outcome when their respective partisan agents do not hold government power.