Clients sometimes despair and ask me: “what’s the one thing that no-one has thought of to get faster growth?” Yesterday’s decisively “meh” GDP data focuses the mind further on such questions. The question is ridiculous. South Africa knows its problems and the answers; the issue is implementation and managing expectations during long, complex work to get things moving. South Africa’s growth problem is an onion with layers of constraining issues. Just because we focus on some key issues and solve a few (such as the Operation Vulindlela agenda) does not mean others aren’t ready to become binding after an initial growth spurt. The president has seemingly alighted on three crucial legacy projects for the next 18 months (assuming someone other than “none of the above” is elected at the ANC conference in December next year): state-owned enterprise (SOE) unbundling, skills and employment institutional reform, and criminal justice big bangs. Each of these is quite orthogonal to the others and interlaced and causal to other constraints and processes unfolding and driving growth. They are clever, efficient choices. I can handwave wildly about the broad mess and incoherence that is industrial policy (on which, amusingly, even the ANC in parliament at the recent committee meetings on the industrial development “strategy” seemed to have a rather sceptical eye), but this isn’t really a single thing. No, the one thing is procurement policy. It feels like we have been stuck on the hamster wheel of doom on this issue for too long. Rules land, someone sues, the courts step in and drafting starts afresh, with each cycle burning years of delivery we do not have. Procurement is where the growth agenda is tested, and we have barely started. About R1-trillion a year, about 15% of GDP, now runs through a framework whose parent act sat before the Constitutional Court in May, with judgment still pending. If that act is struck down the new regulations fall with it. Ten years have gone like this. That leaves a question for the government that has nothing to do with what the rules say: did anyone design them to hold in reality and, given the reaction, function through the courts? It seems not, looking at the latest version of the regulations. The government has recast R1-trillion of yearly spend without publishing the socioeconomic impact assessment that section 63(3)(c) of the act obliges it to produce. Its 100% ownership gate departs from the broad-based BEE codes’ own 51% test, so a supplier must now pass two definitions at once, which is an equal treatment challenge waiting to be filed in court. Meanwhile, data-driven contract monitoring is assumed in municipalities that still run paper files. Treating legal durability as a design constraint rather than a risk to manage later would change all of that: publish the impact assessment before commencement rather than after, then align the thresholds with codes the courts have already accepted, and tie each new duty to the capability that makes it possible so that the weakest bodies are not set up to fail. The government has recast R1-trillion of yearly spend without publishing the socioeconomic impact assessment that section 63(3)(c) of the act obliges it to produce. Its 100% ownership gate departs from the broad-based BEE codes’ own 51% test, so a supplier must now pass two definitions at once, which is an equal treatment challenge waiting to be filed in court. The wider issue is how entities function within this in crisis mode and trying to turn around network industries in particular. Eskom and Transnet have been making the same complaint for five years and both told parliament in March 2022 that procurement rules were stopping them from operating commercially. Back came SCM Instruction 3 of 2021/22 and the Treasury called the problem solved. The complaint carried on anyway, because a plant manager waiting on a turbine part does not care what an instruction note permits in principle, only how long the sign-off takes. The tempting answer is to lift commercial SOEs out of the Public Finance Management Act (PFMA) altogether. It should be resisted because weak boards, monopoly-minded executives, and absent consequence hollowed out these firms and compounds existing implicit subsidies from quasigovernment balance sheet support that cheapens debt for these entities. Everything turns instead on the exceptions. Deviations, expansions, section 79 departures and section 92 ministerial exemptions are case-by-case negotiations, with no service or turnaround standard, no deadline and no appeal. A departure properly granted now can still surface as an audit finding in three years. Everything ends up grinding to a halt and taking far too long. The Treasury should focus on whether the exemption machinery can be made fast, standard and publicly transparent, with turnaround targets, standing delegations for maintenance and spares, and a guarantee that a lawful departure stays lawful. This will be increasingly important as the private and public sectors merge in joint ventures around private sector participation projects of all kinds, in particular at Transnet. The question then is if we can get off this hamster wheel on the broad structure of procurement rules and the machine it operates under. The Treasury has been steadily abandoning the historic “PFMA as religion” mindset that pervaded it for much of the past 30 years and is now more open-minded and flexible, willing to find something that works while preserving the principles behind such frameworks. Maybe this is yet another thing to add to the plate of issues, though raising it leads to madnesses such as pushing local content too fast and beyond what can be delivered. But the increasing digitalisation of the state that the Treasury is spearheading ― if combined with deeper reforms and leadership capacity in the office of the chief procurement officer ― could finally get us off the wheel and deal with a crucial, if rather “unsexy” reform. • Attard Montalto leads on political economy, markets and the just energy transition at Krutham, a South African research-led consulting company.Business Day