zaf-df-pases-1995-2024-v1.1
Post-Apartheid Socio-Economic Series 1995-2024
PASES 1995-2024
| Name | Country code |
|---|---|
| South Africa | zaf |
Household Survey
The Post-Apartheid Socio-Economic Series (PASES) is a stacked series of 28 nationally-representative cross-sectional household surveys covering every year between 1995-2024, with the exception of 2000 and 2001. The source data consists of data from the October Household Surveys conducted every year by Statistics South Africa between 1995-19991 and the General Household Surveys, conducted every year by StatsSA since 2002. The current PASES includes data up to 2024. The surveys collect both household and person information and the series currently consists of just over 640 000 households and their 2.4 million members. A subset of variables from the survey data has been included and harmonised across time. Along with some new and derived variables, there are currently 119 variables in the dataset covering social, demographic and economic outcomes. The data are released in two files, a main file and an extra income variables file along with the syntax files used to create the data.
Sample survey data
Households and individuals
v.1: Edited, anonymised data for public distribution
2026-07-13
Changes from PASES 1995-2024 version 1 to version 1.1 are listed below. For further information, users are referred to the do files for the code to create these variables.
Remittances (hh_rem)
The hh_rem variable did not capture missing responses properly and instead these were reflected as fully non-missing. This has been fixed so now this variable has some missing values.
Data for 1997-2008 has been added to the hh_rem variable. This variable now covers 1997-2024. Because of differences in the questions asked between the OHS and the GHS, and the GHS 2002-2008, and 2009 onwards, the derivation of this hh_rem variable is not consistent over the time series.
hh_rem was derived from two variables in the OHS. hh_rem was coded to 1 if an individual received income from remittances, and/or if not-employed individuals said that they were supported by someone outside the household.
Additionally, hh_rem was derived from two variables in the GHS 2002-2008. hh_rem was coded to 1 if not-employed individuals said that they were supported by someone outside the household, and/or the main source of household income was remittances.
hh_rem was derived from one variable in the GHS 2009 onwards. hh_rem was coded to 1 if the household received remittances as an income source.
Pensions (hh_pension)
The hh_pension variable did not capture missing responses properly and instead these were reflected as fully non-missing. This has been fixed so now this variable has some missing values.
Data for 1997-2009 has been added to the hh_pension variable. This variable now covers 1997-2024. Because of differences in the questions asked between the OHS and the GHS, and the GHS 2002-2009, and 2010 onwards, the derivation of this hh_pension variable is not consistent over the time-series. hh_pension was derived from one variable in the OHS. hh_pension was coded to 1 if an individual received private pension income. Additionally, hh_pension was derived from one variable in the GHS 2002-2008. hh_pension was coded to 1 if not-employed individuals said that they were supported by private pension income. hh_pension was derived from one variable in the GHS 2009 onwards. hh_pension was coded to 1 if the household received private pensions as an income source.
Note that the hh_pension variable looks too high in 2009. This is the result of the original variable in the 2009 GHS having far more positive answers for pension income than in other years.
Grant income
The hh_grant variable has been recoded to a dummy only taking values 0 and 1.
Earnings imputations
Earnings imputations have been added for the OHS years 1995, 1997-1999.
In version one employed bracket respondents for 2022-2024 with code 88 did not have their earnings imputed. This has been corrected in version 1.1 of the data.
Household income variables
Household income variables have been made available for OHS years 1997-1998. These are the only years which had all four components required to create the household income variables. However the questions in the source surveys were not always the same between the OHS and the GHS and the final household income variables appear noisier for the OHS (see Remittances and Pensions above)
Missingness and zero incomes are more common in the 1997 and 1998 variables. Handling of missing data for the original variables used to derive income components appears to have varied across these years. For example, the hhinc_imputed variable has no missing data for household income in 1997 (itself concerning) but 1998 has 6.4% . Zero incomes are also highest in these years, at 7 and 9% respectively.
PASES covers a variety of socio-economic and demographic outcomes in one dataset at both the person and household level. The focus of this first version of PASES is on labour market variables, mainly labour market status, and the harmonising, weighting and imputing of earnings information. The data also includes a consolidated set of variables on demographic and household characteristics, geography, education, household income, social grants, fertility, and transport.
The data is nationally representative.
The lowest level of geographic aggregation of the data is Province
The Statistics SA surveys covered all de jure household members (usual residents) of households in the nine provinces of South Africa, and residents in workers' hostels. The surveys did not cover collective living quarters such as student hostels, old age homes, hospitals, prisons, and military barracks.
| Name | Affiliation |
|---|---|
| Andrew Kerr | University of Cape Town |
| Amy Thornton | University of Cape Town |
| Martin Wittenberg | University of Cape Town |
| Start | End |
|---|---|
| 1995 | 2024 |
PASES is based on anonymised publicly available survey data from Statistics South Africa.
| Name | Affiliation | URL | |
|---|---|---|---|
| Manager, DataFirst | University of Cape Town | http://www.datafirst.uct.ac.za | info@data1st.org |
Public access data for use under a Creative Commons CC-BY (Attribution-only) License
Kerr, A., Thornton, A., and Wittenberg, M. Post-Apartheid Socio-Economic Series 1995-2024 [dataset]. Version 1.1. Cape Town: DataFirst [producer and distributor], 2026. DOI: https://doi.org/10.25828/49GK-P872
| Name | URL | |
|---|---|---|
| DataFirst Helpdesk | support@data1st.org | http://www.support.data1st.org |
ddi-zaf-datafirst-survey-year-v
| Name | Affiliation | Role |
|---|---|---|
| DataFirst | University of Cape Town | Metadata creator |
2026-07-17
Version 2