Projects Under Development 

1. Health Pandemic, Job Losses and Recovery in South Africa (with  Kilumelume, M)

The COVID-19 pandemic has caused job losses in most economic sectors worldwide. During the surge of the pandemic, many governments implemented mitigation measures, such as social distancing, stay-home orders (SHOs), and lockdowns, to contain its spread. These measures often altered work conditions, leading certain firms to adopt Work from Home (WFH) practices. Consequently, this created substantial impacts on labor demand and job opportunities. Using unique (confidential) administrative firm job records that span 2018-2022 in South Africa, we document three main findings. First, the pandemic induced a significant reduction in employment opportunities in South Africa, with notable variations between economic sectors. Second, the duration of jobs, both permanent and casual, decreased substantially during the nine months of intense COVID-19 activity. Third, by the end of 2022, jobs and employment opportunities had yet to recover to levels observed before the pandemic. Specifically, employment and working days remain lower, approximately 2 and 4 percent, respectively, compared to the average levels two years before the outbreak. Finally, we document the heterogeneous effects of economic recovery two years after the pandemic. These results underscore the potential short-term adverse effects of the pandemic on employment gaps in key economic sectors in the country. Full Working Paper coming soon!

2. The Price of Political Alignment: Evidence from Firm Audits in South Africa (with Kunal, S  & , Chingwere, J)

How political alignment with the ruling party influences firm audit outcomes in South Africa? This paper investigates whether political alignment with the ruling party influences the intensity and outcomes of firm audits in South Africa. Leveraging a regression discontinuity design based on close election results from 2014, we examine how firms are treated in municipalities narrowly won versus narrowly lost by the African National Congress (ANC). We combine detailed audit data with corporate income tax (CIT) returns to assess whether audit frequency, enforcement outcomes, and firm-reported liabilities differ along political lines. Our findings show a marked reduction in the audit amounts, number of audit cases and compliant taxpayers in ANC-aligned municipalities. These effects extend to tax reporting behavior: firms in aligned areas report significantly lower corporate tax liabilities despite exhibiting higher reported sales. This pattern suggests that political alignment dampens enforcement without corresponding declines in firm performance. Our results provide new evidence of regulatory forbearance shaped by political incentives, raising broader concerns about fairness, institutional credibility, and the erosion of tax morale in politically connected jurisdictions. Full Working Paper coming soon!

3. Can Road Upgrades Improve Local Tax Revenues  and Jobs? Evidence from Rwanda (A. Musonera and D. Overbeck)

This paper investigates whether road infrastructure can increase local tax collection in a low-income country setting. Drawing on a novel dataset which combines information on the timing and exact location of all road pavings throughout Rwanda with a decade of administrative tax records, we estimate significant and sizable positive effect of road pavings on surrounding municipalities' tax revenues. These effects increase over time and are particularly pronounced for tax types which are collected and used by local municipalities. Additional analyses point to business creation, employment growth as well as increases in land values around paved highways as important channels. In contrast, we estimate precise null-effects for central government's tax revenues. Overall, these additional tax revenues make up for at least 4\% of the total cost of the road paving cost. Our evidence therefore suggests that while road investments may not yield substantial central fiscal returns, they play a key role in stimulating local economic activity and enhancing municipal revenue capacity. Full Working Paper, coming soon!

4. How do Xenophobic Attacks Affect Demand for South Africa Exports? (with Kazima, M and Mensah, J)

Xenophobic violence poses a recurring threat to the social stability and economic growth of affected countries. In this paper, we present new empirical evidence of the economic implications of xenophobic attacks on the demand for South African exports. To achieve this, we leverage the unique administrative customs records of South African exporting firms matched with Xenowatch data to identify the spike events. Based on difference-in-differences approach, our findings show that xenophobic attacks have a negative and significant impact on exports. This impact is most significant in the first six months following the event and is salient in the wholesale and retail sector. The decline in exports following xenophobic attacks highlights the vulnerability of the country’s trade dynamics to domestic instability and emphasises the material consequences of xenophobia on economic performance. The study contributes to the literature on the economic impacts of social unrest, suggesting new insights into how domestic disturbances can disrupt external trade relations in emerging markets. Full Working Paper, coming soon!

5. Automation, Firms, and Workers: Evidence from South Africa (with Kilumelume, M, Mensah, J and Sen, K)

We examine the effects of automation on firm performance and labor market in the context of an emerging economy. To do this, we leverage unique administrative data on the universe of manufacturing firms bin South Africa to identify causal evidence of firm-level outcomes from automation adoption. In the event-study design, we derive direct effects of automation on automating firms, the spillover effects on non-automating firms, and employment within the same industry and location. Specifically, our results show that automated firms doubled production and Value Addition (VA) in the first four years of automation. We also find that the adoption of automation resulted in a robust increase in VA per worker, Markup and total factor productivity (TFP) of approximately 86- 97% within first-four years of automation. The study further shows that automation increases employment and average wages and reduces employee turnover in automating firms. However, we find no significant firm-level spillovers on non-automating firms in the same industry and location. Full Working Paper, coming soon!

6.  Education Returns to  a Public Health Insurance Program: Evidence from Rwanda (with Fadzayi  Chingwere) Current working version!

We study the spillover effects of large-scale public health insurance program on early learning outcomes in Rwanda. The policy, which expands affordable healthcare services especially among rural and low-income families, was introduced in different district hospitals at different time. Leveraging detailed administrative data covering universe of primary school national exam test scores in Rwanda and a difference-in-difference approach that exploits within-region variation in exposure to the reform across district hospitals, we document robust evidence that early life exposure to public health insurance leads to a 0.064 standard deviation (approximately 3.4 percent) increase in student performance. We show that the increase in student test scores is mainly driven by the ease of household health costs, substantial increase in prenatal care services and improvement in household education expenditure.

Working Papers

Selected Articles