Licensing & Access¶
This site is documentation: it hosts no data files and links to none. The files, and the scripts that build them, are supplied on request where the terms below allow it.
There is no single licence for the archive. The project's own work (transcriptions it paid for or made, corrections ledgers, crosswalks, panels built from public sources, code and this documentation) can be released openly. Much of what the archive holds belongs to someone else: a licensed database, a university archive, a data service or a publisher. Their terms decide what can be shared, and the project's licence does not override them.
Terms by dataset¶
Open means the project's version may be shared under the licence named, with the citation given on the dataset page. On request means it is shared with researchers, for research use, under the holder's terms. Not shared means users must get it from the holder.
| Dataset | Underlying source and holder | Terms that bind us | The project's version |
|---|---|---|---|
| Agricultural census | Census volumes of the Department of Statistics / Central Statistical Service (state copyright) | Statistical facts from a public government publication; cite the volume and table | The project's transcriptions and corrections ledgers: open, CC BY 4.0. The 1961–83 panel, which includes the corrections made by Christopher Boone's research assistants, is the project's; it is not yet shared and will be released under CC BY 4.0 once the 1974-shock paper has been submitted |
| Mining census, Manufacturing census | Department of Mines, Department of Statistics / CSS publications (state copyright) | As above | Open, CC BY 4.0 (transcriptions and panels) |
| Elections and the electoral crosswalks | Government Gazette returns; B. M. Schoeman's compilations (copyright of the book's publisher); Delimitation Commission reports and maps (state copyright) | Election results are facts; the compilation itself is not reproduced | Open, CC BY 4.0 for the transcribed returns, the 1977 Gazette transcription, the crosswalks and the division traits. Cite the Gazette or Schoeman beside this archive |
| Forced removals | Surplus People Project, Forced Removals in South Africa (1983), copyright SPP | Facts from a published report; the volumes are not reproduced | Open, CC BY 4.0 for the event file and district panel, citing the SPP volume and page |
| Labour-agent licence areas, 1918–85 | Government Gazette notices (state copyright) | As for other government publications | Open, CC BY 4.0 |
| Open/closed district classification | NRC board memorandum, 1974 (TEBA archive), as reproduced in Crush (1993), Figure 1 | Digitized from the published figure, not from the memorandum | Open, CC BY 4.0, citing Crush (1993). The memorandum itself is TEBA material (below) |
| Eiselen Line coding, district roster, geographic controls | Eiselen's 1955 definition (Goldin 1984); project coding | — | Open, CC BY 4.0 |
| Maize suitability by district | FAO/IIASA Global Agro-Ecological Zones (GAEZ) raster, averaged over GADM district polygons | FAO/IIASA user agreement: non-commercial use with acknowledgement; no resale. Which GAEZ version the file comes from is not recorded. GADM's boundaries may not be redistributed | On request, non-commercial, with the FAO/IIASA citation. Not CC BY |
| Rainfall district panel | Daily station records of the South African Weather Service, compiled by Lynch (2003) for the Water Research Commission (WRC Report 1156/1/03) | Station data are SAWS records; we hold no licence for redistribution, and none has been asked for | Not shared, neither the station data nor the district-year aggregates, until the WRC and SAWS confirm what may be passed on. Users can obtain the database from the WRC and rebuild the panel with the builder |
| TEBA mine recruiting and everything else from the TEBA archive | Records of TEBA and its predecessors (NRC, WNLA), University of Johannesburg Library (UJLIC) | Signed UJLIC permission of 15 August 2013: research use only, no commercial or publication rights, copyright remains with TEBA, content not to be altered, and "should the research be published at a later date, permission from UJLIC and TEBA will be obtained" | Not shared publicly. Not in any public deposit or on this site. See TEBA material |
| Population census microdata (1960–2001), Current Population Survey 1980, the 1993 PSLSD household survey | Statistics South Africa and its predecessors, distributed by DataFirst (University of Cape Town); the PSLSD by SALDRU and the World Bank | DataFirst's user agreement, accepted at download, which does not allow passing the files on | Microdata not shared: request them from DataFirst. Per-district aggregates the project computed are open, CC BY 4.0, except the 1960 Afrikaans home-language shares, which stay on request until DataFirst confirms they may be published |
| Manpower Survey, 1965–94 | Department of Labour / Department of Manpower and the Central Statistical Service. DataFirst distributes the survey as a public-use file (version 1.2, 2015; login required) | DataFirst's user agreement binds users of version 1.2. The project's file is an earlier copy of the data tapes, supplied by Andrew Kerr (DataFirst, University of Cape Town) in January 2013 with no restrictions on use, ahead of the public release (it still carries variable G9, which version 1.2 deletes) | Tape file not shared: users should download version 1.2 from DataFirst. The summary panel and benchmarks built from it are open, CC BY 4.0 |
| Strike panel | Strike reports in Work in Progress (Southern African Research Service, Braamfontein); coding by the project's research assistants | Facts from a published periodical; the reports are not reproduced | Open, CC BY 4.0 for the incident file and its coding. The firm matches are project work and open; any column holding share prices or returns is not (below) |
| Share prices and market values | Datastream (now LSEG), Compustat Global (S&P), Global Financial Data | Subscription licences that forbid redistribution | Not shared. Users need their own subscription. Results computed from them (abnormal returns averaged over events) are reported in the book |
| Company boards, McGregor's Who Owns Whom | McGregor's Who Owns Whom (now McGregor BFA), copyright of the publisher; McGregor BFA electronic extracts | The printed directories are copyright works; the electronic extracts are licensed | Not shared while the extracts or a transcription of the directories are the content. Firm-level indicators the project derives (for example, a strike firm's parent) can be shared once checked |
| Orbis, UNIDO, World Bank Enterprise Surveys | Bureau van Dijk; UNIDO; World Bank | Licensed or registration-only | Not shared |
| Macro series from CEIC and Datastream | CEIC Data; Datastream | Subscription licences | Not shared |
| FRED, World Bank and ILO series | Federal Reserve Bank of St Louis (with series-specific owners), World Bank, ILO | Mostly open with attribution; some FRED series carry their owners' copyright | Users should download them from the source |
| Industrial council wage agreements | Collected wage-rate volumes (SALDRU series); the agreements were gazetted | Facts from published agreements; the volumes are not reproduced | On request until the extraction is checked; then open, CC BY 4.0 |
| Input-output tables | Central Statistical Service / Statistics South Africa | Official statistics, reusable with acknowledgement | Users should take them from Statistics South Africa; the project's conversions are on request |
| SAIRR survey extracts | South African Institute of Race Relations, A Survey of Race Relations (copyright SAIRR) | Copyright publication | On request, research use; not checked |
| Public sector employment (Standish 1987; Brand 1984) | Standish's thesis and Brand's paper for the Second Carnegie Inquiry (copyright of the authors) | Copyright works | On request, research use, citing the original |
| Firm management scores ("AMP" tables) | Described in the archive's notes as World Management Survey data; the source readme has not been checked | The survey's data terms; not confirmed | Not shared: users should register with the World Management Survey |
| HSRC opinion surveys, 1981 and 1983 | Human Sciences Research Council | HSRC data terms; not yet checked | Not shared until checked |
| Published literature | The publishers | Copyright | Not distributed. The Literature pages carry reading notes, bibliographic details and short quotations only |
The 2013 transcriptions were commissioned by Laurence Wilse-Samson, from oDesk freelancers and from Rao's Infocube (the latter under a purchase order through Columbia University), and he holds the rights in them. The 1961–83 agricultural panel is likewise his; Christopher Boone's research assistants corrected it for the 1974-shock paper.
Licence text for the open datasets¶
Each open dataset ships with a short notice, adapted from the following:
[Dataset name]. Compiled by Laurence Wilse-Samson for the South African Apartheid Economy Data Archive. The compilation, transcription and coding are released under the Creative Commons Attribution 4.0 International licence (CC BY 4.0), https://creativecommons.org/licenses/by/4.0/. The underlying figures come from [source: title, volume, table or page], published by [holder]; that publication is not reproduced and remains under its holder's copyright. Cite both this archive and the original source. The data are provided as is, with the known errors and open items listed at https://laurencehw.github.io/sa-apartheid-economy-data/provenance/.
For a dataset shared on request, the notice instead names the holder's terms, states that the copy is for the requesting researcher's non-commercial research only, and asks that it not be passed on. TEBA-derived material is not supplied on request (see below).
TEBA material¶
The terms come from the permission form signed at the University of Johannesburg archive (UJLIC) on 15 August 2013, read from a photograph of the signed form: research use only, no commercial or publication rights, copyright remains with TEBA, content not to be altered, and "should the research be published at a later date, permission from UJLIC and TEBA will be obtained." That permission has not yet been requested.
The following therefore stay out of every public deposit and off this site:
- the TEBA recruiting panels (
teba_panel_*) and the UJ annual-report extractions; - the NRC wage-file extractions and the board memoranda;
mine_location_indicator.csv(NRC/TEBA member-mine records);country_panel.csv(TEBA national recruiting totals);- the recruiting-station lists and
station_district_crosswalk_v4.csv.
Columns derived from TEBA records inside joined panels (recruiting counts, station distances) carry the same terms. The 2013 permission covers the project's own research, not passing TEBA-derived files to other researchers: anyone who needs them should apply to UJLIC and TEBA, and the project supplies its derived files only to a researcher who holds that permission in writing.
Replication packages¶
Each paper's replication package is deposited on Zenodo, with a DOI, at acceptance. Until then, cite this site and the date consulted.
Each package's Data Availability Statement and LICENSE-data list the restricted inputs it needs and how to obtain them. For the 1974-shock paper (Boone and Wilse-Samson) these are the TEBA-derived files teba_panel_v5.csv, mine_location_indicator.csv and country_panel.csv, and the DataFirst-derived afrikaans_lang_proxy_1960.csv; Boone's private district_panel.dta, used only for an optional cross-check, is not shipped. The driver, run_all.py, stops if a required file is missing, so a full run needs the TEBA-derived files, and so permission from UJLIC and TEBA first.
Requesting data¶
Write to Laurence Wilse-Samson at NYU Wagner. Name the dataset and the intended use, and confirm that the use is non-commercial research. TEBA-derived files are supplied only to a researcher who holds written permission from UJLIC and TEBA. Material whose holder forbids redistribution (Datastream, Compustat, CEIC, McGregor BFA extracts, DataFirst microdata) cannot be supplied, nor can the rainfall data until the WRC and SAWS confirm what may be passed on; the request will be answered with where to get it.