This paper investigates how subsidiary autonomy affects the possibility and intensity of innovation in subsidiaries of multinational enterprises (MNEs). Subsidiaries are increasingly identified as vehicles for cross-border transfer of new competences and as an important source of knowledge contributing to innovation. The reason is that knowledge absorbed by the subsidiary from its external and its internal networks are conducive to subsidiary innovation. A key challenge for subsidiaries is the need to balance internal networks within the MNE network and external networks in the local host environment. Both have been shown to be relevant sources of knowledge. If not, subsidiaries lose opportunities to update new knowledge or create a new idea from different knowledge sources. Thus, how subsidiaries balance internal and external knowledge sources in generating innovations at the subsidiary level. We argue that the level of decision-making autonomy of subsidiary is a crucial missing mechanism in our understanding of subsidiary innovation. The question of how much decision-making autonomy subsidiaries should have is a core issue in the management of headquarter-subsidiary relationships. Using two divergent theoretical perspectives, we hypothesize that decision-making autonomy allows subsidiaries to adapt and use local knowledge and business networks required to develop new products and processes. We also hypothesize that too high levels of autonomy may have negative effects on innovation levels in the subsidiary because of a reduction in supports from headquarters, leading us to predict an inverse U shaped relationship between subsidiary autonomy and innovation. We test our hypotheses in a multi-country and multi-industry database based on survey evidence of 775 subsidiaries located in five Central and Eastern European countries (CEEs) that serve headquarters in 44 different nation states. The empirical results provide strong support for a non-linear relationship between decision-making autonomy and product, process, marketing, and organizational innovation, respectively, as well as between decision-making autonomy and innovation intensity.