Smallholder agricultural value chains remain as one of the most widely supported market-based strategies for achieving rural development. This thesis goes beyond measuring the monetary benefits from value chain participation, seeking to conceptualise how trade-offs and heterogeneous rural development outcomes may emerge from smallholder value chains. Drawing on a range of disciplinary perspectives, this thesis utilises narrative, spatial modelling, econometric, and psychometric methodologies across two case study regions in Uganda and Indonesia. In doing so, the thesis seeks to highlight methodological blind spots in empirical value chain research, and to provide internally valid research findings to support value chain policy in each case study region. Importantly, the thesis develops concepts and methods that are externally valuable for broader value chain research and design.

The thesis begins with a scoping literature review to articulate the domains of which smallholder value chains act as rural development tools. These domains include the ability of smallholders to participate in smallholder value chains (inclusion), the benefits these smallholders and rural communities receive from value chains (value), and the extent in which value chains can scale and remain financially sustainable (scalability). The review then explores to what extent trade-offs between these three domains emerge for smallholder-targeted value chain programs. From a sample of 344 case studies emerges a strong pattern of trade-offs and heterogenous outcomes which inform the remaining focus of this thesis: patterns of value chains exclusion, heterogeneous interhousehold and intra-household experiences from value chain participation, and tradeoffs between agricultural development and broader policy objectives such as sustainability.

The second chapter seeks to understand why some of these trade-offs between outcomes may emerge using an agent-based linear programming model (ABM) of upland landscapes of Bandung, Indonesia. The model simulates value chain interventions to understand how these interventions may perform in respect to poverty alleviation, inclusion, and broader environmental concerns. The modelling results show the extent in which transaction and investment costs impede inclusion of smallholders in coffee agroforestry value chains, and highlights how environmental objectives of agroforestry value chain interventions may fail due to the complex and diverse land use decisions made by heterogenous smallholders.

The scoping review also highlights methodological limitations of value chain inclusion and how exclusion is currently conceptualised and empirically identified. Chapter four introduces empirical methodologies to identify and measure value chain inclusion opportunities, which separate out value chain inclusion/exclusion outcomes derived from structural barriers from the choices and preferences within a smallholder’s control. These inclusion methodologies combine insights from the inequality in opportunity literature with stated preference experiments, and was tested with a case study of coffee farmers in Kapchorwa, Uganda.

By assessing the inequality associated with expected participation outcomes based on circumstances alone, we show that coffee value chains are moderately exclusive. This finding holds for binary (extensive) and continuous (intensive) measures of value chain participation, measured at both the household and individual level. Spatial factors, such as altitude, and disagreements over financial norms within the household are particularly associated with value chain exclusion.

The final chapter explores how these disagreements over financial norms may also result in heterogeneous intra-household outcomes from value chain payments. This chapter presents psychometric models derived from social psychology literature to understand how disagreements in household financial norms create intentions for hidden financial behaviour amongst coffee growing households in Uganda. The results show that the failure to resolve differences in household norms regarding expenditure patterns is significantly associated with the intention for private expenditure for women, but not so for men. Private consumption strategies representing just one approach chosen by women where the benefits of hiding consumption (e.g. being able to consume in ways that better reflect their personal preferences) outweigh the costs (which may include psychosocial costs such as guilt and the risks of being ‘found out’ by their partner). Equitable bargaining power appears to play an important, albeit partial, role in achieving greater household cooperation.

Throughout, the thesis calls into questions whether value chains can act as effective rural development tools. There appears to be a common basic assumption that the private sector can, through self-interested actions, generate positive social outcomes associated with rural development. This is a considerable assumption and relies heavily on the (unlikely) case that incentives for private (business) outcomes happen to generate the same organisational structures as concerns for the rural communities. Value chains can still act as important moderators for welfare gains for some, but likely not all, cohorts.

The four thesis chapters are all currently submitted to international peer review journals. The scoping review presented in Chapter Two is currently under review at World Development. Chapter Three has been submitted to Land Use Policy, Chapter Four to World Development, and Chapter Five to Feminist Economics. In addition, all chapters have been presented at conferences in Australia and Internationally, including the International Conference of Agricultural Economists (2024), the Australian Society of Agricultural and Resource Economics Conferences (2022, 2023, and 2024) and the University of New England Research Pathways Conference (2024).

Global prevalence rates suggest that three to four males are diagnosed as autistic for each female (Loomes et al., 2017; Zeidan et al., 2022). To explain this, there is some suggestion that autistic females have a different clinical presentation of their autism symptoms, and that this is influenced by the degree that they ‘camouflage’ their autistic behaviour (Allely, 2019b; Hull, Petrides, et al., 2020b). Camouflaging has been defined as “the employment of specific behavioural and cognitive strategies by autistic people to adapt to or cope within the predominantly non-autistic social world” (Cook, Hull, et al., 2021, p. 1). However, some researchers have argued that there is a lack of clarity in camouflaging studies, such as inconsistency in the terms and definitions used in describing this behaviour (Fombonne, 2020). The present research reviewed the extent of these conceptual issues, plus some measurement and methodological limitations that may have resulted from a lack of clarity (Chapter Two). To address this lack of clarity, conceptual analysis was employed in two Chapters. This involved synthesising findings from autism research with concepts from two domains of the wider literature. First, in Chapter Three, concepts and theory on a behavioural consequence of stigma (i.e., negative stereotypes, attitudes, and beliefs held about autistic people; Link & Phelan, 2001), known as concealment, were reviewed. From this synthesis, a definition of autism concealment was given, and its empirically supported behavioural characteristics were reviewed. A novel theory was developed to explain the correlates of this behaviour. It was argued that the concept of camouflaging could largely be understood by reference to this framework, which was more informed by the wider literature than camouflaging studies. However, some aspects of camouflaging, argued to resemble forms of social learning, were unexplained by this framework. To resolve this, another conceptual analysis was conducted (Chapter Four), in which concepts from the wider literature on recovery and compensation, two forms of adaptive brain plasticity, were extended to the autistic population by framing them as neural mechanisms underlying biobehavioural resilience. This concept was supported by a scoping review of studies of the brain-behaviour relationships of autistic people, and a conceptual framework was constructed to integrate several modulators of this process. Finally, to address some of the measurement issues of camouflaging research, a new psychometric scale was developed to measure autism concealment: the Autism Concealment Scale. Its development process was assisted by a content validation study, and its psychometric properties were largely supported in a second exploratory study (Chapter Five). Limitations to the replicability of its factor structure, and some aspects of its validity, warrant future study. Strengths, some other limitations, and various implications of the present research, such as the impacts of these concepts on the reliability of autism diagnosis, were discussed (in each Chapter, and summarised in Chapter Six).