JAKE MORRIS: Despite the data, SA still needs manufacturing

SA’s manufacturing sector is not doing well. After an impressive run of 65% growth in real value added from 1993 to 2008, the sector remained flat for a decade until 2019, and since then has declined 6%, Stats SA reports. 

Services are flying. Finance, real estate and business services have grown 40% since 2008, while personal services (mainly private healthcare) have grown 24%. Since 2008, manufacturing’s share of GDP has dropped from 16% to 13%, whereas the above services’ share of GDP has grown from 36% to 44%. 

Further arguments in favour of services include that modern technology has made services more tradable — while tourism still can’t be exported online, financial and business services can. Another is that some digitalised services (think Google, Netflix, Spotify) lend themselves to global-scale economies that raise productivity and economic value like never before.

Add to this the concern that manufacturing jobs are especially vulnerable to automation, and we might ask ourselves: should SA forget about manufacturing and chart a services-based growth path instead?

The answer is no. As Cambridge economists Jostein Hauge and Ha-Joon Chang tell us, factories made the modern world and they will keep remaking it. Here’s why. 

It’s unclear how much manufacturing has genuinely declined relative to services. We should be cautious with the data for the following reasons.

Standard Industrial Classification (SIC) codes that categorise firms into sectors do so based on the type of work done by most employees. Consider a manufacturer with 100 employees, 70 directly involved in manufacturing and 30 involved in related services such as research & development (R&D), marketing, security and logistics. If the firm doubles in size and outsources these noncore activities, the associated value and employment are reclassified as services. This makes it appear that manufacturing activity has dropped relative to services, whereas the opposite has happened.

Productivity gains

Studies show that often a decline of manufacturing relative to services has more to do with higher productivity gains in manufacturing, which result in lower prices and therefore total value, than with less manufacturing being done. This is because despite some digital services being far more productive than before, many core services do not lend themselves to productivity gains. It is difficult for a doctor to see a patient for less time or for a restaurant to spend less time preparing a meal without compromising the quality of the services.

One indicator that SA manufacturing may not have declined quite as much as the data suggests is that since 2019 it has had the highest employment growth rate (6.2%) of any private sector of the economy other than healthcare. And this is ignoring the data’s bias against outsourced manufacturing services. 

Whatever the numbers, the international evidence clearly shows that every country (with more than a few million people and without enormous oil reserves) that has developed from low- to high-income status has done so through manufacturing. This is how “industrialised country” and “developed country” came to mean the same thing.

Economic growth has generally slowed when manufacturing has shrunk and services have risen. SA is no exception. Between 1993 and 2008, when manufacturing held constant at 16% of GDP, our average GDP growth rate was 4%. From 2009 to 2023, when manufacturing’s share of GDP fell to 13%, average GDP growth was only 1% (or 2% if the 2009 and 2020 crisis years are left out). The causal relationships are undoubtedly more complex, but the correlation is striking.

So why is manufacturing so central to economic growth? Manufacturing is the primary driver of productivity growth over time. It is also the primary driver of technological development and spillovers into the rest of the economy. This leads on to the third reason: its multiplier effect. Aside from technology transfer, services rely on manufacturers as customers. When manufacturing grows, services grow too.

But surely services also create output in other sectors, including manufacturing? Not to the same extent. Studies show manufacturing creates two to three times as much output in other sectors compared with services. In Singapore, every 100 new manufacturing jobs are associated with 27 new non-manufacturing jobs, whereas every 100 new services jobs are associated with only three manufacturing jobs. 

In fact, some services are so tightly linked to manufacturing that they should arguably be classified as manufacturing activities. Consider industrial R&D, product design and industrial engineering, which exist solely for the purpose of manufacturing. 

Cautionary tale

While some services have become far more tradable, as a category it still pales in comparison with manufacturing. The proportion of world trade accounted for by services has barely moved from 20% in 1980 to 23% in 2022. 

Even India, the poster child of outsourced information communication technology (ICT) services, offers a cautionary tale. It has a trade deficit because its trade surplus in services makes up only a fraction of its trade deficit in goods. This year India’s finance minister proclaimed that despite its enviable success in services it needs to figure out how to manufacture more goods.

If this explains the historical promise of a manufacturing-led growth path, what does the threat of automation replacing labour mean for the future?

Despite the hype, there is no clear evidence that manufacturing jobs in developing economies are being or could be replaced at scale by fourth industrial revolution technologies such as robotics and 3D printing. Some studies have found that the threat to jobs in developing countries from automation is as low as 2%-8% of current jobs, with the risk greater in higher-income countries.

Others have found that manufacturers investing in new technologies have actually grown employment as they have become more competitive. And even if the hype about automation replacing jobs en masse were true, it’s not clear that manufacturing would be more vulnerable than services. Raise your hand if you’ve recently reached a chatbot instead of a customer service agent. 

Services are critical to economic development, but they do not trump manufacturing and there is no evidence they will any time soon. We need both. A debate on whether to rely on one or the other is about as useful as arguing over whether Ronaldo or Messi was a better footballer. Both were indispensable and arguably needed each other to prosper.

If we want to prosper we have no choice but to create a globally competitive manufacturing sector. The more useful debate is how to do that. What seems clear is that SA manufacturing needs help, and the government has a major role to play.

A step change in basic service delivery is crucial, but won’t cut it. This is the time for aggressive, smart and selective industrial policy, coupled with competent execution. The ball is in trade, industry & competition minister Parks Tau’s court. 

• Morris is COO at manufacturing value chain management consulting company BMA.