Official source
Source domain: gtac.gov.za
Collected on 31 July 2026
Always confirm details on the official source before acting.
What this means
This document is a slide-style research presentation titled “Automation and inequality within firms: Evidence from South Africa” by Rafael de la Vega, Tim Kӧhler, Antonio Martins-Neto, and Siphelele Ngidi. It presents a study using matched employer-employee tax microdata for South Africa’s formal manufacturing sector (2014–2021) to estimate causal effects of automation adoption on employment, earnings, and within-firm wage inequality. The results reported are that automation adoption reduces employment, shows no average effect on earnings, but increases within-firm earnings inequality in the short term, driven by changes in employment composition (especially affecting middle-wage workers in routine manual tasks).
Facts
Places
City of Tshwane, Gauteng, City of Cape Town, Western Cape
Reference
CIT-IRP5
Key Takeaways
- Title: Automation and inequality within firms: Evidence from South Africa
- Authors: Rafael de la Vega; Tim Kӧhler; Antonio Martins-Neto; Siphelele Ngidi
- Event: GTAC Public Economics Conference, Pretoria
- Date: July 2026
- Study period/data: 2014–2021
- Sector: South Africa’s formal manufacturing sector
- Data: matched employer-employee tax admin microdata (universe of firms and workers)
- Firm data: CIT-IRP5 panel; job data: IRP5 employee tax certificates
- Automation measure: imports of focal capital goods (6-digit HS codes) tied to automating manufacturing work
- Method: staggered semi-parametric Difference-in-Differences (Callaway & Sant, 2021) using automation “spike” (first year of highest firm-specific import value)
- Outcomes: employment; mean earnings; within-firm inequality (GE with alpha 0, 1, 2)
- Reported findings: automation reduces employment; no evidence of effect on average earnings; positive short-term effect on within-firm earnings inequality
- Inequality mechanism: employment composition effects rather than earnings changes; concentrated on middle-wage workers in routine manual tasks
- Focal capital goods categories include: automatic data processing machinery, automatic machine tools, dedicated machinery, automatic regulating instruments, automatic conveyors, automatic welding machines, 3-D printers, industrial robots, weaving and knitting machines, electronic calculating machines
- Rafael de la Vega
- Tim Kӧhler
- Antonio Martins-Neto
- Siphelele Ngidi
- UNU-MERIT
- Development Policy Research Unit, University of Cape Town
- World Bank
- Southern Centre for Inequality Studies, University of the Witwatersrand
- GTAC Public Economics Conference
- University of Cape Town
- University of the Witwatersrand
- Callaway & Sant (2021)
- Kerr (2020)
- Bhorat et al., 2023