Currently, due to their level of proximity and periodic transfers of competences from the national government, local governments are responsible for many services consumed by citizens. Their competences are growing and nowadays include water supply, promotion of regional development, solid waste collection, emergency management, health care, education, land use management, among many others. To cope with all these challenges, local governments changed from a single multi-purpose organization to a complex network of relationships with other public bodies, private agents, and non-profit organizations. In this complex network we witness different levels and combinations of autonomy and control that local governments allow/impose on their agencies/partners/contractors. Local governments in Portugal no longer assume the provision of all public services through in-house bureaucracies. They combine different strategies of coordination using both market competition and network collaboration based on trust as alternative mechanisms. The path of evolution began with simple hierarchic organizations composed by municipal services, then it moved to municipalized services (a slightly more autonomous and flexible configuration) followed by enabling legislation allowing local governments to create municipal corporations (Tavares & Camões, 2007), which in many cases replaced former arrangements. By the end of the 1990s, local governments began contracting with external actors to provide public services. Their option was to contract-out public services making use of market price mechanisms and market competitiveness. In other cases, local governments preferred less competitive partnerships with other governments or with non-profit organizations (mostly on social services). A survey conducted among local governments allows us to identify twelve organizational configuration alternatives to deliver public services: municipal services; municipalized services; municipal enterprises; inter-municipal enterprises; municipal commercial corporations; public commercial corporations; mixed commercial corporations; contracting-out of private agents; partnerships with non-profit actors; local governments associations; and metropolitan associations. Using coordination and control mechanisms as a criteria (authority, competition and collaboration), we seek to identify all the alternatives presented in three governance mechanisms (hierarchy, market, and networks). 2 Municipal Service Delivery: The Role of Transaction Costs in the Choicebetween Alternative Governance Mechanisms. EGPA Annual Conference, Malta, 2-5 September 2009 Next, we employ a transaction cost framework to derive a set of hypotheses explaining the choice of governance mechanisms to provide public services. In order to do so, we develop four sets of independent variables concerning service characteristics, community characteristics, financial situation, and administrative and political stability. Services with high levels of specificity and difficult measurability involve high levels of transactions costs, making them less likely to be externalized (Coase, 1937; Williamson, 1975; Nelson, 1997; Ferris & Graddy, 1997). Second, we argue that local governments operating in highly complex environments require more flexible governance solutions (Burns & Stalker, 1961) and, in contrast, mechanisms of governance based on hierarchy are more likely to be adopted in environments characterized by lower levels of complexity in service provision (Weber, 1947; Alexander, 1995; Blau & Meyer, 1971). Third, we argue that different degrees of financial capacity drive local governments to choose different mechanisms of governance (Brown, Potoski, & Slyke, 2006). Finally, based upon work by Hood (1998), Frant (1996), and Clingermayer and Feiock (1997), we suggest that political instability can put pressure on politicians to improve efficiency levels through the use of external agents. We test our hypotheses with a multinomial logit regression model using survey data collected from 101 Portuguese local governments and archival data available from the National Bureau of Statistics. The Portuguese Science and Technology Foundation (Fundação para a Ciência e Tecnologia – FCT) [grant number PEst-OE/CJP/UI0758/2011-2012] provided financial support for this research. Any opinions, findings, and conclusions or recommendations expressed in this material are those of the authors and do not necessarily reflect the views of the Fundação de Ciência e Tecnologia(FCT).