OpenUCT is the open access institutional repository of the University of Cape Town (UCT). It preserves and makes UCT scholarly outputs digitally and freely available, including theses and dissertations, journal articles, book chapters, technical and research reports, as well as open educational resources.

 

Recent Submissions

Item
Open Access
An early-warning system for emerging market crises: a machine-learning approach
(2026) Barow, Dylan; Sayed, Ayesha
This research report develops an interpretable machine-learning early-warning system for financial crises in emerging market economies. While traditional early-warning system models identify key macro-financial predictors, their reliance on linear specifications often fails to capture the non-linear dynamics and cross-country heterogeneity inherent in modern crises. In contrast, recent machine-learning approaches offer greater predictive flexibility, yet are frequently criticised for lacking policy relevance due to their “black-box” opacity. Although the trigger events surrounding a crisis can materialise rapidly, the predictive content of crises lies primarily in the gradual accumulation of macro-financial vulnerabilities over multiple years, a process that is well captured at quarterly frequency and that aligns with the medium-term horizon over which macroprudential policy operates. Using a quarterly panel of 25 emerging market economies from 1995 to 2024, this study constructs indicators for currency, banking, and sovereign debt crises. Gradient-boosting models are estimated within a walk-forward expanding-window framework to generate probabilistic forecasts at four- and twelve-quarter horizons. Performance is assessed using discrimination, calibration, and policy-usefulness metrics that reflect the asymmetric costs of missed crises relative to false alarms. The results demonstrate that ensemble methods, particularly gradient-boosting classifiers, consistently outperform linear benchmarks, with predictive power improving consistently at longer horizons, supporting the focus on medium-term horizons for policy surveillance. Shapley Additive exPlanations (SHAP)-based interpretability reveals that crisis risk is shaped by non-linear threshold effects in domestic credit conditions, external imbalances, and global financial conditions. Overall, the findings suggest that machine-learning frameworks can significantly enhance macroprudential surveillance by combining improved predictive accuracy with the transparency required for timely policy intervention.
Item
Open Access
Leveraging artificial intelligence for enhanced echocardiography: using machine learning for Congenital Heart Disease
(2026) Dzonzi, Brenald; Ngoepe, Malebogo; Mishra, Amit
Coarctation of the aorta (CoA) is a congenital cardiovascular defect characterised by a focal narrowing of the descending aorta, resulting in elevated cardiac afterload and often necessitating lifelong haemodynamic surveillance. Although computational fluid dynamics (CFD) enables non-invasive assessment of pressure gradients and flow fields, its reliance on labour-intensive image segmentation, mesh generation, and substantial compute time limits its applicability for rapid parameter studies and routine clinical use, particularly in resource-constrained settings. Neural network surrogates offer a promising alternative by learning input–output mappings from CFD data at a fraction of the inference cost, yet the relative merits of physics informed versus purely data-driven architectures remain insufficiently characterised for patient-specific cardiovascular geometries. This dissertation benchmarks physics-informed neural networks (PINNs) against data-only deep neural networks (DNNs) as mesh-free surrogates for haemodynamic prediction in CoA, with evaluation on a patient-specific geometry from the PROTEA project at the University of Cape Town. A structured benchmark ladder of increasing physical complexity is employed: (i) analytic ordinary and partial differential equations (spring–mass system, two-dimensional Poisson equation) providing exact reference solutions; (ii) canonical laminar flows (planar Poiseuille and three-dimensional pipe flow); (iii) a parametric stenosis series with varying constriction severity; and (iv) a clinically derived CoA geometry with pulsatile inflow. Model families—supervised DNNs with optional random Fourier feature embeddings, physics-only PINNs, hybrid PINNs combining data and residual losses, and a signed-distance-function-augmented variant—are implemented in PyTorch and trained using automatic mixed precision, adaptive loss weighting, and reproducible multi-seed protocols. OpenFOAM supplies the ground-truth CFD datasets and boundary conditions throughout. Results demonstrate that physics constraints confer a substantial advantage under extreme data scarcity (the PINN achieves an 8.5× lower extrapolation error than the DNN at ten labelled points on the Poisson benchmark), but this benefit diminishes as labelled data increase. On the patient-specific CoA geometry, the baseline DNN and Fourier-enhanced DNN satisfy the adopted pressure-drop acceptance criterion (below 1% mean relative error) at both 50% and 100% data fractions, whereas the hybrid PINN exhibits systematic gradient imbalance between its data and physics objectives. Fourier feature embeddings deliver comparable accuracy to the baseline at one to two orders of magnitude lower training cost, representing the most favourable cost–accuracy trade-off observed. All models are validated against CFD rather than in-vivo measurements; clinical deployment would require subsequent catheterisation or four-dimensional flow MRI validation and multi-patient generalisation. This work establishes reproducible technical foundations—encompassing data generation, training, and post-processing pipelines—for neural-network-based haemodynamic surrogates in congenital heart disease, contributing toward the long-term goal of accessible, rapid cardiovascular assessment.
Item
Open Access
“Roads to Riches”: The Infrastructure-Property Development Nexus in South Africa
(2026) Jelley, Bruce; Albertus, Rene
This minor dissertation investigates the nexus between public-sector infrastructure spending and new property formation, a proxy for economic growth, at the regional level in South Africa. This relationship is critical to sustainable economic growth and to achieving the United Nations' Sustainable Development Goals. While the global literature highlights infrastructure as a catalyst for growth, empirical findings have become increasingly contested, particularly at the regional level, where spatial disparities, baseline infrastructure gaps, and fiscal inefficiencies prevail. In South Africa, inadequate infrastructure and declining service delivery have constrained property development, raising questions about whether state-led investment meaningfully stimulates construction activity and economic growth. The study adopts a positivist quantitative approach, applying an Autoregressive Distributed Lag (ARDL) model to quarterly provincial and national data from 2011-2025. This period captures weak and uneven post-2011 growth, technical recessionary episodes, the COVID-19 contraction, and the broader institutional context associated with state capture. Building plan approvals serve as a forward-looking proxy for new property formation, rather than a direct measure of completed construction, with public sector infrastructure expenditure as the primary explanatory variable alongside other determinants of property formation, including prime lending rates, inflation, foreign direct investment, and working-age population. The methodology accounts for both short-run dynamics and long-run equilibrium relationships, supported by unit-root testing, bounds testing, and residual diagnostics to ensure robustness. Findings reveal statistically significant long-run relationships between infrastructure spending and new property formation across Gauteng, KwaZulu-Natal, and the Western Cape, with elasticities ranging from 0.44% to 0.53% for every 1% increase in public sector infrastructure spending. These elasticities are economically meaningful but moderate, suggesting that infrastructure expenditure is an important enabling condition rather than a complete explanation of property-sector growth. Short-run results show more nuanced regional differences, including counterintuitive positive effects of higher lending rates in KwaZulu-Natal. Inflationary pressures dampen activity in certain provinces, while foreign direct investment shows limited explanatory power, contrary to classical FDI-growth theory.
Item
Open Access
The economic burden of diagnostic pathways: An analysis of patients costs at the Rapid Access Diagnostic Lymph Adenopathy Clinic (RADLAC) at Groote Schuur Hospital in Cape Town, South Africa
(2026) Byaruhanga, Christine; Cunnama, Lucy; Paidamoyo, Bodzo
Globally, cancer is one of the leading causes of mortality and morbidity, with the greatest burden in low- and middle-income countries where approximately 65% -70% of deaths occur. Cancer is the non-communicable disease most likely to lead to catastrophic health expenditure. This can occur at any phase along the continuum of cancer care. There exists a gap in the literature on expenditures incurred prior to receiving a diagnosis. Delays in receiving a diagnosis may lead to adverse health outcomes along the cancer care continuum; moreover, in some cases, these delays can cause financial distress for the individual and household. The Rapid Access Diagnostic LymphAdenopathy Clinic (RADLAC) was established to enable rapid diagnosis of patients with lymphadenopathy. The opening of this clinic provided a chance to understand the types, magnitude, and drivers of cancer-related costs for patients seeking a diagnosis as well as their households. The aim of this study was to analyse patient costs associated with seeking a diagnosis at the RADLAC at Groote Schuur Hospital, a tertiary academic hospital in Cape Town, South Africa. The objectives of the study were to calculate the direct and indirect costs related to diagnosis for patients attending the RADLAC. The study also measured the intangible dimension of pain. Coping strategies used by patients to cover costs were also investigated. The study was a cross-sectional patient cost analysis. Data were collected from consenting newly referred and previously referred adult patients. A structured questionnaire collecting direct (medical and non-medical) and indirect costs was used to collect data via in-person interviews. Pain was measured using the Brief Pain Inventory (BPI), which has been validated for South African use, to capture the intangible dimension associated with lymphadenopathy (LAP). Data was collected in REDCap and analysed using Stata18. Patient costs were calculated up to the point of diagnosis. We found that patients incurred significant direct costs prior to attending the clinic, with out-of-pocket (OOP) payments spent on general practitioners and transport costs making up the majority of these expenses. For some individuals, monthly transport costs to RADLAC accounted for 15% of their monthly expenditure. Patients also lost earnings whilst seeking a diagnosis with 29% reporting lost earnings from attending the RADLAC. Most patients had informal caregivers who lost a significant proportion of their daily wages each time they had to accompany the patient to seek health care services. Of the patients who had someone accompany them to the clinic, 44% of them reported lost earnings on that day Patients coped with the costs of health care using a multitude of strategies. Almost 80% of low-income households reduced spending on other household categories to pay for health care costs. It was observed that many individuals also borrowed from informal lenders who charged interest rates as high as 50%. Pain caused by lymphadenopathy was a common symptom reported by patients. Most patients reported that the pain due to lymphadenopathy and/or its accompanying symptoms interfered with their daily functioning. In addition, pain also affected their sleep and their enjoyment of life. In some cases, patients had to incur additional direct medical costs to manage lymphadenopathy-associated pain, i.e., additional trips to the pharmacy for pain medication. South Africa has pledged to achieve Universal Health Coverage (UHC) through the implementation of the National Health Insurance (NHI). Comprehensive financial risk protection is critical to achieving UHC and relates to the pursuit of Sustainable Development Goal 3. Direct and indirect costs of seeking a diagnosis influence health-seeking behaviours and influence health outcomes. Despite care being free at the point of service at primary health care facilities in South Africa, patients incur substantial direct non-medical costs in the form of transport costs. Indirect costs in the form of foregone wages are also incurred each time care is sought, and these losses account for a significant proportion of patients' monthly income. Further costs may be incurred when accessing secondary or tertiary facilities rather than primary health facilities. Comprehensive health, social, and economic policies are needed to ensure that patients are adequately protected from the costs of illness, particularly the costs associated with cancer diagnosis and care.
Item
Open Access
Social safety nets and household vulnerability to income shocks and food insecurity among poor households in Malawi
(2026) Chibade, Evans; Alhassan, Abdul Latif
This study assesses the effectiveness of formal and informal social safety nets (SSNs) in reducing household vulnerability to income shocks and enhancing food security among poor households in Malawi. Utilizing data from the Fifth Integrated Household Survey, the research employs a set of econometric models, including the multivariate probit, Poisson regression, Tobit and ordered logit, to analyze the impact of social safety nets on three key outcomes: participation in income-generating activities, household food consumption expenditure, and multidimensional food security indicators, including Household Dietary Diversity Score, Reduced Coping Strategies Index, Food Consumption Score, and Food Insecurity Experience Scale. The findings reveal that government cash transfers are effective in stimulating investment in productive income-generating activities like livestock and non-farm businesses, thereby enhancing dietary diversity and food consumption scores. Both formal and informal social safety nets play a crucial role in consumption smoothing and reducing reliance on negative coping strategies. A critical finding is that households acting as providers of support within informal networks exhibit significantly greater resilience and food security than those solely receiving aid. However, the transformative potential of SSNs is constrained by structural factors: Urban residence, mobile phone ownership, and diversified livelihoods are key determinants of resilience. The study concludes that while SSNs are essential for immediate welfare and protection, achieving sustainable poverty reduction requires integrating them with complementary "cash-plus" interventions, such as access to extension services and market linkages, to address the underlying causes of vulnerability.