The Case for Intelligent Industrial Policy

Earlier this year, German Economy Minister Peter Altmaier unveiled his “National Industrial Strategy 2030,” which aims to protect German firms against state-subsidized Chinese competitors. The strategy identifies key industrial sectors that will receive special government support, calls for establishing production of electric-car batteries in Europe, and advocates mergers to achieve economies of scale.

The planned measures are controversial. Lars Feld of the German Council of Economic Experts calls the strategy an aberration and has accused Altmaier of central planning. Yet this is not an ideological debate, as Feld suggests, but rather a question of whether such an industrial policy might work. And although not all aspects of Altmaier’s plan are convincing, there is a strong case for government support to sectors – including the automotive industry – that will increasingly rely on artificial intelligence (AI).

True, national industrial policies generally have a bad reputation among economists, mainly because governments have typically used such policies to support “losers,” thereby keeping uncompetitive firms in the market. The argument in favour of such policies, especially in developing countries, was that infant industries needed protection from foreign competition in order to grow and mature. But the World Bank concluded long ago that these policies had failed, turning its back on import-substitution programs in the 1960s and 1970s.

The emergence of strategic trade policy in the 1980s provided a theoretical foundation for an active industrial policy. In a perfectly competitive world market, the optimal export policy is not free trade, but a sufficiently small export tariff. But the situation changes if firms have market power and compete strategically with one another, as in the case of European aircraft manufacturer Airbus and its American rival Boeing. Here, subsidizing Airbus can result in that firm gaining market share from Boeing.

Altmaier and his French counterpart Bruno Le Maire recently used this rationale to push for a Franco-German rail-industry merger between Alstom and Siemens. They argued that the tie-up would produce a European champion capable of taking on the Chinese rail giant CRRC.

But their arguments were unconvincing. Whereas subsidizing Airbus created an additional competitor, the proposed Alstom-Siemens merger would have reduced the number of European rail companies. Moreover, as EU Competition Commissioner Margrethe Vestager pointed out when blocking the merger, Alstom and Siemens rarely compete with CRRC in third countries, because the Chinese company mainly focuses on its home market. So the merger probably would not have enabled Alstom-Siemens to capture market share from CRRC.

The case for government intervention is much stronger in sectors that have economies of scale, and where “learning by doing” creates knowledge that remains even after a state subsidy ends. This applies in particular to the many sectors that will soon be powered by AI. The more data a firm or a sector produces, the more it learns and the better the AI algorithm becomes.

Because of its large size, China has a comparative advantage in these sectors. The Chinese authorities have recognized this and have cleverly used state subsidies to promote AI and support domestic firms. To help counter this challenge, Germany and Europe should respond with their own subsidies to knowledge-based sectors.

Europe’s automotive industry is an obvious candidate for such support. This is why Altmaier’s plan to develop European battery production for electric cars make sense, and may even accelerate an industrial renaissance across the continent.

Producing electric-car batteries in Europe would attract more automotive companies and lead to lower car prices, because producers would need to import far fewer batteries from Asia. Such growth may in turn create a self-reinforcing agglomeration effect, as additional car companies move production to Europe in order to be near other auto firms and their suppliers. Paul Krugman and Anthony J. Venables outlined these backward and forward linkages in a well-known article more than two decades ago.

A sizeable electric-car industry in Europe would increase demand for labor and cause real per capita incomes to rise. And the more important battery cells are to the overall value of electric cars, the stronger the agglomeration effect will be.

This would be a momentous shift. China currently controls the entire value chain of electric cars, including the supply of cobalt, an essential raw material in battery manufacturing. As a result, China produces 69% of the world’s electric-car battery cells, the United States 15% (at Tesla’s “Gigafactory” in Nevada), and Europe just 4%. Nonetheless, Europe could still develop battery production by recycling old electronics, developing new cobalt-saving processes, and discovering alternative mineral deposits.

Some might argue that Europe should let its auto industry gradually migrate to China, which can make cars more cheaply. After all, consumers want to buy reasonable-quality cars at the lowest possible price. If the Chinese can produce them more efficiently, the argument goes, then let them do it. But, quite apart from the huge political and economic backlash such a decision would generate in Europe, this view is too simplistic. AI will be central to the future of the car industry, and advances in this field will have positive spillover effects in many other sectors.

If Europe gives up on its automotive industry, it will lose knowledge and future growth. The German government is therefore right to support the country’s carmakers. Far from being a relic of socialist central planning, industrial policy may sometimes be the intelligent choice.