South Africa has a long history of oppression and apartheid which have led to great inequalities, despite South Africa's classification as an upper-middle income country (World Bank, 2008). 26 years after the fall of apartheid, the systematic racial segregation practiced under apartheid, in conjunction with an overtly white supremacist ideology still has a profound impact on South Africa's society as well as its education system (Spaull, 2012). Howie (2012) explains how South Africa struggles with a widening performance gap between rich and poor students and high levels of drop out, particularly among black Africans. Spaull's (2012) analysis of SACMEQ III data shows that students in the 25% wealthiest schools are far more likely to have their own textbook, receive homework frequently and experience less teacher absenteeism compared to the poorest 75% of schools. The distribution of resources and capital still privileges white South Africans, according to Nattrass and Seekings (2001) and Spaull (2012) and essentially divides the country and the education system into two nations. Several authors point to problematic accountability relationships and a lack of trust and capacity in the South African education system as key issues in the lack of improvement of learning outcomes. Spaull (2001) for example explains how the national, provincial and local levels of government are not held accountable for their use of public resources, and how there are few (if any) tangible consequences for non-performance or to address the high rates of teacher absenteeism and low rates of curriculum coverage. Eddy Spicer, Ehren et al's (2016) systematic review also points to lack of teacher accountability in South Africa as school-based registers of teachers' attendance are not checked and national government fails to sanction teachers who are often absent. Lack of trust inhibits the implementation of effective assessment and inspection systems as teacher unions, for example, reject inspections of teachers and block the publication of assessment data, while lack of capacity subsequently prevents district managers, area managers, principals, heads of department and teachers to effectively use the data that is available (see Eddy Spicer, Ehren et al, 2016; Howie, 2012). This 'vicious' cycle of distrust, lack of accountability and lack of capacity renders the system powerless to improve and creates a series of 'binding constraints' (e.g. weak institutional functionality, undue union influence) that need to be addressed in order to improve learning outcomes, according to Van der Berg et al (2016). These examples highlight the need to understand the intricate relations between accountability, capacity and trust and how these relations produce (or fail to produce) a pattern of change in learning outcomes over time and create a divided unequal system. We aim to study these relations in South Africa's public primary education system in three phases: 1) Social network analysis of the (accountability and trust) relations and flow of resources in a low and high performing school in quintile 1 and 5 (20% of schools in most deprived and wealthy areas), 2) A focus group in which we will map the causal loops which describe how trust, capacity and accountability interconnect through a series of balancing and/or reinforcing feedback loops and how these loops have (in the past) produced (or failed to produce) a pattern of change in learning outcomes and have created large inequity in the system, and 3) Collecting quantitative (assessment and questionnaire data) to test the causal loops via longitudinal path models and multiple-indicator multiple-cause (MIMIC) models.