Ethiopia is an agrarian country where agriculture dominates the economy, and thus agriculture is considered as an engine of growth by the government. Seed as one of the agricultural technologies, in fact, a carrier of many technologies, is critical to increasing production, but the use of quality seed from formal sources in Ethiopia is very limited. The current Ethiopian government has focused on agricultural development and has developed different policies both for agriculture in general and for the seed sector in particular. Following the developmental state approach, the government intensified its involvement in the seed sector to enhance agricultural development. Despite the policies and efforts of the government, a shortage of seed, a mismatch between demand and supply, the carryover of seed despite not satisfying the demand of farmers, and poor seed quality have been persistent challenges to the Ethiopian seed sector. Many studies have identified technical gaps that limit the development of the seed sector, and some of the studies have also discussed the extent to which policy responds to existing problems, and the extent to which what is in the policy documents is implemented. However, the causes of these ‘gaps’ are seldom discussed. The lack of such knowledge limits the understanding of the challenges, making it difficult to properly support the seed sector. For these reasons, this research has gone beyond the mere identification of ‘gaps’, aiming to analyse how actors and institutions influence seed policy making and implementation in Ethiopia. The goal of this research is twofold: to narrow the knowledge gap about policy making and implementation in the Ethiopian seed sector, and to contribute to the debate concerning how to make the seed sector function better. The central research question is: how did actors and institutions influence the formulation and implementation of seed policies in Ethiopia from 2008 to 2016? The empirical research to answer this overall research question addresses two processes: policy making and policy implementation. These include the process of revising the 2000 Ethiopian seed law and the process of implementing direct seed marketing. By analysing these two processes, the thesis unravels how actors and associated institutions have influenced seed policy making and implementation in Ethiopia. The major sources of data were interviews of actors in the seed sector, and desk research of different reports. Guided by theoretical concepts, the research used qualitative methods to generate and analyse data. Given the complexity of societal phenomenon, several analytical lenses have been used to examine the data in this research. In order to explain how actors negotiate the content of a policy document, including defining the problem and solution, the concept of discourse analysis is used, focusing on frame, the rounds model, and the policy arena. Similarly, to explain the process of implementing the existing policy and the outcome, the concepts of multi-level perspective on transition, transition management, non-decision making, and institutional lock-in are used. While using these analytical lenses to explain seed policy making and implementation, the concept of institutions has remained a central concept. Chapter 2 analyses the negotiation process, looking into the topics of seed sector governance and variety registration. The analysis reveals that different policy arenas provide opportunities for different actors to place their preferred policy options on the table, and to get them incorporated into the draft working document. While this is a positive step towards a deliberative policy making, the final decision is made by the executive branch of the government. Such a process can be explained by two informal institutions. These are the loose connection between the drafting arenas and the decision-making arenas, and the blurred separation of power between the executive and the legislature. At the Council of Ministers (CoM), where the critical decisions are made, the ministry presents its perspective, particularly on issues where disagreement exists between the ministry and other actors. The council uses the content of the draft and the justification of the ministry for endorsing the draft policy document. Moreover, the parliament can change the content of the draft policy document only if the ministry agrees with the change, regardless of the arguments and justifications provided by other stakeholders. Thus, the inputs of stakeholders are considered as long as the ministry agrees with the suggestions, and the policy decision remains in the hands of the ministry. Chapter 3 presents the different frames used by different actors to describe the problem of seed quality. While government officials attribute the problem of seed quality to the lack of alignment between the seed sector governance and the regional government structure, experts and bureaucrats attribute the problem to the lack of coordination at national level. As a result, they respectively suggest the decentralization and centralization of seed sector governance. These frames are embedded in the overall interest and strategy of the actors promoting the frames. The centralization frame reflects the interest of experts and bureaucrats to have a say with regards to the seed sector. They have lost this power because of the federal structure that was established formally in 1995. On the contrary, the decentralization frame is embedded in the government’s aim to implement the constitution that established the federal structure in 1995. Despite the fact that the process of revising the seed law took about four years, these actors could not agree on either of the options or find an alternative. This shows a lack of deliberation and reflexivity during the process of revising the seed law, reflecting the fact that seed policy discussion has been part of a larger debate about (de)centralization in Ethiopia since 1991. Thus, in addition to the issue of seed quality, the frames of centralization and decentralization are shaped by the old (unitary) and the new (federal) institutions of the Ethiopian government system. Chapter 4 focuses on the process of introducing and expanding direct seed marketing (DSM) in Ethiopia. Despite the fact that seed marketing is included in the policies on paper, the seed of major food crops is distributed through government channels resulting in inefficiency of seed distribution. The regional seed core groups introduced DSM in 2011, and by 2016 about one-third of the hybrid maize seed, the main seed marketed in Ethiopia, in Amhara, Oromia and the Southern Nations, Nationalities and Peoples’ region (SNNPR), was sold through DSM. The presence of actors outside the seed distribution system was instrumental for introducing the concept of DSM. To start the piloting of this existing policy, the core group needed to get approval from the heads of the bureaus of agriculture (BoAs). However, such approval was not required for other new ideas, like establishing an independent regulatory body, showing how the informal institutions guide what has to be approved by bureau heads, regardless of the formal policy. In addition to the demonstrated potential of DSM to overcome the problem of seed distribution inefficiency, strategic management of the stakeholders' process was critical in expanding the area under the pilot. Many actors, including the executives, supported the expansion of DSM to many areas. Despite the expansion of DSM, its demonstrated potential to overcome the problem of seed supply inefficiency, the support it received from the government officials, and the general policy of market-based approach, the government has not endorsed the use of DSM beyond the pilot. Chapter 5 points out that the government excluded the issue of seed marketing from the seed regulation enacted in 2016, showing that the government has no intention to make seed marketing one of the seed delivery channels in the near future. The major reasons for this are: bureaucrats do not want to contribute to the decision making of DSM because they assume that the government has a strong political interest to remain in seed distribution; bureaucrats need the seed distribution system to achieve the targets set by the government; there is a symbiotic relationship between actors, the extension service as well as seed producers, and the seed distribution system, and so actors want to maintain the distribution system Such institutionalized thinking and practices have created an institutional lock-in that prevents bureaucrats from presenting the recommendation to government officials, thereby leading to non-decision about the future of DSM. Chapter 6 summarizes the action of actors in affecting policy making and implementation as influenced by two conflicting sets of institutions. The first set relates to market-based thinking versus centralized planning as leading principles for economic development. Both are used as a discourse for promoting economic development and its operationalization, which are shaping how actors view and overcome the problems of the seed sector. This also explains why policies on paper are not implemented and why new initiatives are not formally endorsed. The tension between these divergent institutions has increased because of the dual use of seed by the government: the government has used the seed to both promote economic development and maintain strong political ties with farmers. The second set of conflicting institutions relates to authoritarian versus participatory decision making. On the one hand, is the government practice of authoritative decision-making, where only the input of stakeholders is considered when it fits in with the existing policy direction of the executives. On the other hand, it is common practice to organize stakeholders to contribute to policy making and implementation. The practice of considering the policy input of others only when it fits in with the policy direction of the decision-makers, creates a sense of being forced to accept, increasing the tension between how the government decides and the role of stakeholders. Given the tension between the conflicting institutions, and circumstances in Ethiopia, this research suggested that choosing one approach over the other will not guarantee the development of the seed sector. There is no guarantee that the outcome of a deliberative policy making process will be a different policy option than the one opted for by one of the actors. However, the co-development of a solution for the shared seed sector problem will guarantee better ownership and thus better implementation than an imposed policy. It is also important to note that deliberative policy making and implementation is not an easy task given the current stakeholders’ landscape and the culture of authoritative decision making. Thus, the change towards deliberative policy making and implementation is not something that emerges overnight: it is a process that matures over time. This calls for the strategic management of a process of change that leads to the transformation of the seed sector into a self-reliant and resilient sector. By identifying the underlying institutions behind the challenges of the seed sector and suggesting options for improvement, this thesis contributes to the debate on how to make the seed sector function better. At a higher level, it also contributes to the debate on policy making and implementation processes in Ethiopia.