In today’s blog post, we repost an unedited version of our Business Day article in which we argue that he department of communications & digital technologies deserves praise for policies that, if implemented, will improve the SA data landscape.
To the cloud, and beyond
Pietman Roos and Daan Steenkamp
On 31 May 2024, the Department of Communications and Digital Technologies (DCDT) published the National Policy on Data and Cloud. It is for the most part a coherent and sensible paper and the DCDT deserves praise for policies that, if implemented, will fundamentally change the South African data landscape for the better. The new policy declares that government shall prioritise cloud services as the primary option for new ICT procurement. This will be a boon for the local data centre and cloud services industry.
The paper invites stakeholders to give feedback by July. There are five salient policy proposals that are not only worth further consideration, but in fact worthy of their own policy papers. This is not a critique of the paper, but rather a reflection of the complexity of the subject matter.
First, the paper recognizes that government collects vast amounts of data on its citizens in various formats, but more could be done to digitize data assets. For example, some information may be stored in hard copy paper format and other information is stored digitally. That obviously leads to inefficiencies, data gaps and potential for data loss. Too true. Ask anyone who lost a marriage certificate.
The Policy recognizes the need for harmonisation of technical standards across government entities and the need for common data governance mechanisms to enable data integration and system interoperability. But the challenge is immense. Limited data are available in machine readable format, limiting data sharing not just among government entities, but also with the private sector. This will require a significant deployment of technically skilled officials within the State Information Technology Agency (SITA) and each government department, to such an extent that short term advisors will be a necessity to fill the gap.
Second, the paper advocates evidenced-based policy on the back of improved data sharing between government departments. Again, a wonderful idea, but sceptics (and policy nerds) might recall that in the late 2000s the government committed to introduce Regulatory Impact Assessments (RIAs) as part of legislative proposals. It is safe to say that RIAs are unfortunately not part and parcel of South Africa’s policy development framework. Which is to say, the lofty ambition to have evidenced-based policy is likely to suffer the same fate unless there is a fundamental change in approach.
Third, the paper suggests that physical data centres that will provide cloud storage should be encouraged to be self-provisioning for water and electricity, while addressing carbon emission reductions to mitigate against environmental degradation, and that they must not be in areas prone to natural disasters or social disturbances and have a minimum uptime of 99.995%.
Given South Africa’s unreliable public service provision, if such requirements are to be implemented, incentive schemes or other mitigation measures might need to be considered. Alternatively, in the interest of cost efficiency, there should be a graded approach to balance data urgency and accessibility. Even if South Africa’s public data were accessible to all, a downtime of a few hours to access the latest birth records would not be as detrimental to safety and prosperity as say, the unavailability the police’s information system.
Fourth, the paper makes the mature admission that given the complexity and cost of building and operating a data centre, government should leave this to the private sector and rent server space. This is a welcome approach and given the size of the private sector investment and taxpayer outlay, an economic study would be welcome to understand the challenges and opportunities. Although there has been growing investment in local data centres, South Africa still lags other major emerging markets.
The fifth proposal worth highlighting is that the paper states that “The private sector also amasses substantial volumes of data through its transactional relationships with customers, which include citizens, organizations, institutions, and other companies. When made available and utilized, this data can help derive insights that drive innovation, foster creativity in problem-solving, and facilitate the development of new and tradable digital goods and services.”
A wary reading is that this statement is proposing that companies cede their intellectual property. There are of course questions around granularity: could a multinational corporation be compelled to share personal information on its clients to government? Beyond privacy concerns, that obviously creates a risk for commercially sensitive information falling into the hands of competitors.
Government currently collects data by virtue of statute, and data collection that disproportionally impacts rights in comparison to the intended government purpose can be constitutionally challenged. The policy seems to suggest that private entities can be approached to offer up their private data, without any statutory authority. This is no technical quibble here but this strikes at the heart public data and its democratization. The power to collect data should be taken incredibly seriously, and similarly, the citizenry is entitled to access data they effectively pay for by way of their taxes.
An optimistic reading is that government wants to invite private entities into a larger data ecosystem, voluntarily. If that is the case, then more power to the DCDT. However, creating an enabling environment for responsible data sharing for public policy and commercial purposes would require far deeper discussion. The paper recognizes the need for “open standards and open systems, including open-source frameworks”, but in fairness this statement of intent is a starting point for several more rounds of public consultation.
An effective framework to share government data is not a nice-to-have or a first world problem. The Global Data Barometer shows that South Africa is falling behind other emerging markets, like Rwanda, Uganda and Nigeria, in terms of use and sharing of government data. And a World Bank survey found that South Africa does not have “an enabling regulatory environment that promotes the use and reuse of public intent data.”
Maturing the data management infrastructure and practices in the public sector presents a tremendous challenge but must be a national strategic priority for South Africa to achieve the data-led future the National Policy on Data and Cloud imagines.
Dr Steenkamp is CEO at Codera Analytics and a research associate at Stellenbosch University. Roos is an associate with Codera.
