South Africa’s construction regulations only ever seem to move after the rubble is cleared. But look closely at what each revision actually does, and a pattern emerges: the more people die, the more paperwork we impose on the people who weren’t responsible for the collapse in the first place.
Three times now, South Africa has rewritten its Construction Regulations, and three times the sequence has looked the same. A tragedy happens. Ministers appear at a podium and promise never again. A gazette follows. And a few years later, another building comes down.
The 2003 Construction Regulations were born from a chronic problem — poor safety performance and an industry culture where clients and designers simply weren’t at the table when safety decisions got made. That much made sense: the regulations tried to drag stakeholders who had no skin in the game into the room.
Then came Tongaat. On 19 November 2013, part of a mall under construction north of Durban came down, killing two workers and injuring 29 others, after an explosion in a column caused the concrete slab it supported to collapse. A commission of inquiry found the site’s plans had never been approved, that excavation had proceeded without authorisation, and that a “team leader” had ordered props removed from under a beam that wasn’t ready to stand on its own. It was, in other words, a story about unlawful construction and a chain of command that ignored its own risk assessments.
Three months later, the Construction Regulations 2014 were gazetted. And here is where the story gets interesting, because the headline reform to come out of that process wasn’t really aimed at the things that caused Tongaat. It was the forced professional registration of Construction Health & Safety Officers and Managers with a statutory body, the SACPCMP — a body established under a completely different Act, for a completely different purpose (project and construction management), that had never before regulated site-level safety officers.
The Department of Labour and SACPCMP were unambiguous about why this was necessary. As the Council put it at the time, registration was “a critical intervention in terms of enhancing CHS in the sector, and the urgent need to reduce injuries, fatalities, and disease.” That’s a big claim for a paperwork requirement to carry. And the people it fell on didn’t buy it. As one industry commentator, Rudy Maritz, wrote bluntly in response: safety officers “are not responsible for slab collapses, silo collapses, and houses falling apart.” NIOCCSA, representing a chunk of the profession, made the same point in more diplomatic language to the Department directly — that Construction Health and Safety Officers “are currently not engaged in Project Planning, Design Development and Contractor Procurement,” which is precisely where the decisions that cause structural failures actually get made. You cannot professionalise your way out of a collapse by certifying the person whose job is to check for missing guardrails, if the actual failure originated three stages earlier, in a design sign-off or an illegal change of scope nobody re-engineered for.
Worse, the mechanism itself buckled under its own weight almost immediately. Within months of the 2014 regulations landing, SACPCMP admitted its registration process was taking four months longer than promised. By 2017, of more than 14,000 applications submitted since 2013, only around 2,500 had actually been completed, with over 1,700 archived for incomplete paperwork. An intervention sold as urgent public-safety infrastructure spent its first several years functioning mainly as a bottleneck and, by the accounts of the people forced through it, a costly one — annual Continuing Professional Development obligations that could run to thousands of rands, imposed on a workforce earning, on average, R23,000 a month.
None of this means health and safety officers shouldn’t be competent, or that oversight of the profession is a bad idea in principle. It means something more specific and more damning: the 2014 reform mistook the appearance of rigour — certificates, registers, CPD points — for the substance of it, and it did so while the actual causal chain behind the disaster that prompted it went largely untouched.
Which brings us to George. On 6 May 2024, a five-storey development under construction collapsed, killing 34 workers. The investigation that followed didn’t describe a safety-officer failure. It described a building approved as a single storey and silently converted to five, without proper re-engineering; construction that started before the project was even enrolled with the NHBRC; and, per the Minister of Public Works, a legislative framework so outdated that the core National Building Regulations Act hadn’t been meaningfully amended since 1996. That is a story about who gets to change a building’s design without anyone checking, and about a regulator that never noticed. It is, notably, almost exactly the same kind of story Tongaat told eleven years earlier — unapproved plans, work proceeding regardless.
The Draft Construction Regulations 2025 are now open for comment, and the early signs are that the SACPCMP registration net is simply being widened again, catching still more categories of people under the same statutory body whose earlier expansion did so little to prevent the disaster that supposedly justified it. Industry lawyers reviewing the draft have already flagged the same complaint that dogged 2014: legislators “see these big buildings collapsing, and that is who they write the regulations for, and they don’t always realise that it affects a lot of other people.”
If South Africa wants its next Construction Regulations to actually break this cycle, the test should be simple, and it’s the one 2014 failed: does the new rule constrain the people who can authorise an unapproved five-storey conversion, or does it just add another certificate to the wall of someone who was never in the room when that decision got made? Eleven years of evidence suggest we already know which kind of rule we’re better at writing — and it isn’t the kind that saves lives.