In proportional representation (PR) systems, political power is typically shared across several political parties. Many parties are small and focused on specific policies, such as environmental policy and immigration. Do these small parties matter for policy? I provide the first systematic evidence for this by developing the first regression discontinuity design tailored specifically for PR systems. With this method, which can be applied in all countries with PR systems, I estimate the causal effect of party representation on immigration policy, environmental policy and tax policy using data on Swedish municipalities. The results show that party representation has a large effect on the first two policies, but not on the tax policy.