South Africa's failed industrial policy
Rather than import-substitution, South Africa should focus on export promotion through technological innovation
Industrial policy is back in fashion. Over the last few months, prominent US economists have made the case for reconsidering government initiatives aimed at encouraging the growth of specific industries, often through subsidies, tax incentives, and regulation.
Harvard economist Dani Rodrik is one proponent. In a new research paper with Réka Juhász and Nathan Lane, he argues that the benefits of industrial policy have been dismissed too quickly by academic economists. There are at least three good reasons for industrial policy: externalities, coordination failures, and activity-specific public inputs. Positive externalities refer to the unaccounted social benefits arising from economic activities; for example, a firm producing semiconductor inputs can contribute to national security, but these benefits are not fully internalised in the firm’s financial calculations.
Coordination failures occur when one producer’s success depends on others’ activities. An example of a coordination failure is when manufacturers are hesitant to invest in solar panels because there are no suppliers, and suppliers are hesitant to produce solar panels because there are no manufacturers to buy them, leading to a stalemate that government intervention could resolve.
Activity-specific public inputs emphasise that public goods and services, such as infrastructure and education, must often be tailored to specific industries to maximise their impact. An example would be the decision to build science parks near universities to foster innovation in the tech industry. Here, the public good (the science park) is not a generic service but is specifically tailored to suit the needs of tech companies and startups. The close proximity to a university could facilitate knowledge transfer, offer access to a skilled talent pool, and encourage collaborative research, thereby providing a targeted benefit designed to boost a particular sector of the economy.
In all three cases, it is necessary for government to intervene in the market through industrial policy. For that reason, Rodrik and co-authors are adamant that the question is not whether industrial policy should be implemented but how.
And, in truth, most economists would agree with these theoretical benefits. But many worry about its practical implementation; as soon as some politician or bureaucrat must decide which industry to support (or where to build a science park, for example), industrial policy can be easily abused to favour certain regions or people at the cost of everyone else. As the authors themselves argue, industrial policy could ‘open the door to self-interested lobbying and political influence activities, diverting the government into activities that enrich private interests without enlarging the social pie’.
So, it works in theory, but many worry that it doesn’t work in practice. Why, then, are people like Rodrik so optimistic? Rodrik and his co-authors argue that the concerns about political abuse and corruption have been overstated. Recent academic papers, using better methods than earlier studies which relied predominantly on correlations, show that industrial policy can propel economic growth. One very innovative study uses the Napoleonic Blockade of the early nineteenth century to show that regions in the French Empire which became better protected from trade with the British for exogenous reasons during the Napoleonic Wars increased capacity in mechanised cotton spinning to a larger extent than regions which remained more exposed to trade. In the long run, regions with exogenously higher spinning capacity had higher activity in mechanised cotton spinning.
Rodrik and co-authors summarise the main findings from this new branch of research:
A striking finding that emerges across some (though not all) of these papers is just how large and long-lasting the local effects of industrial policy can be. To us, a balanced reading of the emerging literature suggests that it is no longer possible to dismiss industrial policy as inefficient or counter-productive. [T]he market failures that justify its use seem large.
However, there are many types of industrial policy and many motivations for it. Economics blogger Noah Smith lists six reasons to implement such policies: 1) Securing foreign exchange in order to achieve import security of food and energy, 2) catch-up industrialisation in a developing nation, 3) boosting economic growth and/or employment and wages in a developed nation, 4) maintaining a military edge over rival countries, 5) insuring supply chains against the risk of war, and 6) preventing climate change or other negative externalities.
Knowing the motivation for a policy is important because it matters in determining its eventual success. A key finding from Rodrik et al. is that industrial policies should be more outward-looking, focusing on export promotion rather than protectionist measures. Here, history helps: in the twentieth century, many Asian countries built their success on export-oriented policies, whereas many Latin American and African countries attempted import-substituting policies, and failed.
South Africa has several industrial policies in place, although it is difficult to determine from outdated government websites which ones are still relevant. The foreword of a ten-year review of the country’s Industrial Action Plan, written by former minister Rob Davies and published in 2020, provides a useful overview of our focus or, to be more precise, lack of focus. It appears to encompass both import-substitution and export-oriented elements, for example. On the import-substitution side, there is an emphasis on localisation, supplier development, and the need to strengthen State-Owned Entities (SOEs) to provide efficient and competitive basic services. It aims to reduce ‘significant import leakages’ and stresses the need for black economic empowerment. These elements suggest a focus on reducing dependence on imported goods and services by developing local industries – the type of policies Rodrik would warn against.
On the export-oriented side, the foreword discusses the need for a ‘much stronger export effort’ and identifies the African continent and sub-region as an enormous opportunity. Min Davies emphasises supporting existing and new exporters, especially black-owned entrants. This is more indicative of a strategy that is looking to increase its export capabilities and reach.
One method to distil the motivation behind South Africa’s industrial policy – other than reading what the former minister says – is to use natural language processing techniques to investigate what politicians say. The idea is that words that are closely associated with industrial policy would reveal the government’s motivation for the policy. This is not such a bizarre idea. In an earlier paper, Réka Juhász, one of Rodrik’s co-authors, used a similar technique on the Global Trade Alert database, a database that records commercial policy interventions. Words such as ‘export’, ‘boost’, ‘research’, ‘development’, and ‘technology’ are all reasonably associated with industrial policy, she finds. It is clear that industrial policy is to boost exports through technology improvements.
And in South Africa? I asked Phronesis Analytics, a Pretoria University-based start-up, to undertake a topic analysis of political speeches and parliamentary debates in South Africa. They did so for an entire database of speeches and documents from 2012. Four clusters immediately stand out: the first around ‘youth’, the second around ‘special economic zones’, the third around the ‘green economy’, and the fourth around ‘black industrialists and bbbee’. On the balance of evidence, then, it would seem that industrial policy in South Africa is not aimed at improving exports but rather at redistribution, another motivation to be added to Smith’s list. The word ‘technology’ is notably missing from discussions around industrial policy in South Africa.
Whether redistribution and empowerment can be achieved with industrial policy – or whether these worthy goals are not best left to labour or education policies – is an important question. Given the latest evidence, industrial policy not aimed at improving exports through technological innovation seems unlikely to succeed. And even we did target export growth, South Africa might be too small to benefit from an active industrial policy. Rodrik et al. warn, for example, that wealthier countries benefit most from implementing industrial policies.
South Africa is not wrong to want to implement industrial policies. There are good reasons to do so, and some evidence that they can succeed. But the context and, most importantly, the motivation for these policies will determine their success. Instead of following the prevailing fashion of export-focused and tech-driven industrial policies that aim to turbocharge economies, South Africa has opted for political expediency at the expense of economic growth.
We risk wielding the hammer of industrial policy as if every issue is a nail, neglecting the fact that true transformation requires more than just a single tool in the toolbox.
An edited version of this article was first published on News24. The image was created with Midjourney v5.2.
Unfortunately, what is said most or less normally equates to the same negative outcome. Such is our country.
I'm hoping clever people like you give input into topical situations e.g.,
1. What would the effect of AGOA crashing be, and how could SA adapt?
2. SA is standing for Palestine. Considering our trade with Israel, how have we benefitted in the past, and what parts (they need the least) could we lose?