In 2010 the municipality of Norrköping took a decision to implement a social impact bond. This decision was the starting point for a wave of similar decision around the Swedish municipalities. Södertälje took the decision to implement their social impact bond in early 2013. Even though much of the policy was the same as in Norrköping, it didn’t generate the result expected and already in 2018 it was reconstructed and brought back into the daily business. The purpose of this study is to explain the policy decisions taken, both in connection with the introduction of the policy in 2013 and the transformation done in 2018. The analytic framework used to explain these policy changes is based on the Multiple Streams Theory. The study is based on a qualitative research method with information taken from documents and interviews. These are aimed to help us gathering the empirical material needed, which consist of official control documents from the current municipality and interviews with pertinent people who worked within the policy or who have good knowledge of the phenomenon of social impact bonds. Our main conclusions are that the policy decision to implement the social impact bond, did couple the three streams; problem, political and policy, so this decision was done within an open policy window. However, the decision to reconstruct the social impact bond didn’t couple the streams in the same way and because of that a policy window didn't open.