How this chronology was built Methodology & sources

This page sets out what is in the manufacturing chronology, what is not, where the dates and figures come from, and how to get an error fixed. It follows the same standard as the mining chronology, with one significant addition: manufacturing has a boundary problem that mining does not, and that problem is dealt with below.

What this is, and what it is not

The chronology is a secondary synthesis, compiled from the published record; statutes, tariff determinations, commission and board reports, official statistical series, company filings and established scholarship. It involves no original archival research.

Use it to orient yourself and to find out where to look next. Do not cite it for any individual fact. If your work will be fact-checked, peer-reviewed or edited to Wikipedia's sourcing standards, follow the pointers back to the underlying material and cite that instead.

The boundary problem

Mining has natural edges. Manufacturing does not, and any chronology of it embeds a definition whether or not it admits to one. Ours:

• Included: the transformation of materials into goods; metals and engineering, chemicals, motor assembly and components, food and beverage processing, textiles and clothing, paper, cement and building materials, and armaments.
• Excluded: mining itself, construction, and utilities. Electricity appears only where supply failure or pricing directly determined what manufacturers could produce.
• Contested and included: smelting and refining. Ferrochrome, ferromanganese, aluminium and steel sit between mining and manufacturing, and different official series classify them differently. We treat primary metals as manufacturing, which is the Statistics South Africa convention, but readers comparing our chronology against sources that treat beneficiation as mining will find the sectors do not line up.
• Contested and included: Sasol. Coal-to-liquids is extraction, conversion and chemicals at once. We treat it as manufacturing because its industrial significance lies downstream.
• Excluded despite proximity: agriculture, and agro-processing except where a specific plant or policy is at issue.

The overlap with the mining chronology is deliberate. Several events appear in both; the 2008 electricity crisis, COVID-19, the logistics failure; because they bear on both sectors, and the entries are written differently in each to reflect what was at stake for that sector. They are not duplicates.

Scope and selection

Sixty-two entries across 134 years. An event earns a place if it meets at least one of these tests:

• It changed what could be made here. A plant, a technology or a policy that added or removed productive capacity; Iscor, Sasol, the local content programme, the ArcelorMittal longs closure.
• It changed the terms of trade. Tariff decisions, trade agreements, embargoes and preference programmes, which in this sector are as consequential as any investment.
• It changed the labour settlement. The statutes, strikes and union formations that set the cost and organisation of manufacturing work.
• It is a fixed reference point that later accounts consistently anchor to.

Excluded: individual plant openings that did not open a subsector; routine annual tariff adjustments and anti-dumping determinations, of which there are hundreds; ordinary wage rounds; company results, except where they precipitated a closure. Commodity price movements appear only where a specific move had a discrete traceable effect.

Source hierarchy

Date conventions

• Statutes and programmes are filed under enactment, with commencement noted where it differs materially.
• Plants are filed under the year the corporation was established, with first production given in the entry. Iscor therefore appears at 1928 and again at 1934.
• Tariff actions are filed under the date they took effect, not the date they were announced; the reverse of our convention for corporate transactions, because a tariff has no effect until it applies.
• Corporate transactions are filed under announcement, with completion noted separately.
• Strikes are filed under the year they began.

Contested figures and judgements

Three kinds of number in this chronology should be treated as indicative rather than settled.

Sectoral shares

Manufacturing's share of GDP and of employment varies by several percentage points depending on classification, base year and whether primary metals are counted here or under mining. We give approximate figures and describe direction of travel rather than citing precise shares, because precise shares invite comparison across sources that are not comparable.

The 1943 crossover

The point at which manufacturing overtook mining in national output is conventionally dated to the early 1940s, but the exact year depends on the national accounts series used, and the wartime accounts are weaker than what came later. We give it as 1943 and flag that it is a convention.

The liberalisation question

Whether South Africa reduced tariffs too far or too fast after 1994 is the central unresolved argument in the field, and it is a genuine dispute among competent economists rather than a settled matter with a fringe. Entries covering the 1990s state that the argument exists and summarise both positions. We have not adjudicated it, and a reader should not infer a position from the selection.

Category definitions

Each entry carries one thread, assigned by principal consequence rather than by subject matter.

The sixth thread has no counterpart in the mining chronology. It exists because manufacturing's relationship to the outside world; cut off in 1977, reopened in 1990, preferentially admitted in 2000, tariffed again in 2025; is a continuous causal story in this sector and not in that one.

Known gaps

• Firm-level history is thin. The chronology is weighted toward the state and toward large listed companies, because that is where the documentary record is. The small and medium manufacturers who account for most of the sector's firms barely appear.
• Food, beverage and agro-processing are under-represented relative to their share of manufacturing output and employment, because their history is one of gradual consolidation rather than datable events.
• The informal and repair economy is absent entirely. A large amount of material transformation in South Africa happens outside registered manufacturing and is not captured here or in the official series.
• Regional detail is missing. Manufacturing decline has been highly uneven; the former border-industry areas, the East Rand, Newcastle and the Vaal have had very different experiences, and a national chronology flattens that.
• The last two years are provisional and, in the case of the 2025–26 tariff sequence, still moving. Those entries will need revision.
• The armaments industry is treated lightly. Its history is genuinely obscured by decades of secrecy, and the accessible record is not adequate to the sector's industrial significance.

Corrections

If something here is wrong, send the entry, what is wrong with it, and a source. Corrections of fact are made promptly and recorded in the changelog; matters of interpretation and selection are noted but not necessarily changed. Every substantive change increments the version number, and citations should include the version so a reader can tell which text was consulted.

Changelog

Reuse

Published under a Creative Commons Attribution 4.0 licence. You may republish, adapt, translate or build on it, including commercially, provided you attribute it and say if you changed it. The embed code on the chronology page carries the attribution automatically.