CAMBRIDGE – The United States is on a building spree in semiconductors. In early April, Taiwan Semiconductor Manufacturing Company announced plans to set up a third fabrication facility in Arizona to make the world’s most advanced chips, upping its investment in the state to $65 billion. TSMC’s investment is heavily subsidized by the US government under the CHIPS and Science Act, and the company will receive $6.6 billion in grants and is eligible for $5 billion in loans. It can also claim an investment tax credit of up to 25% of its capital expenditures.

This news follows Intel’s recent announcement that it will receive an even larger grant of $8.5 billion from the US government (along with $11 billion in loans on “generous terms”). The CHIPS Act allocated $39 billion for such grants, and additional deals are in the making. According to the White House, nearly $300 billion in manufacturing investments have been committed in the US just over the last two years.

President Joe Biden sees these deals as evidence of a manufacturing renaissance in the US. “Where the hell is it written saying that we’re not going to be the manufacturing capital of the world again?” he asks. His administration may not have much in common with the preceding Trump White House, but it certainly shares a preoccupation with reviving manufacturing.

There are several reasons why manufacturing has once again become the focus of economic policy. For starters, the sector plays a disproportionate role in driving innovation and productivity in the economy, and the pandemic highlighted the risks of far-flung cross-border supply chains. In an age of heightened geopolitical competition, especially vis-à-vis China, US policymakers consider it imperative to manufacture advanced technologies, such as semiconductors, on US soil.

Then there is the goal of creating good jobs. “Sparking a manufacturing, construction, and clean energy renaissance” is at the very top of the administration’s agenda for building a good-jobs economy. On the face of it, this goal makes a lot of sense. Historically, unionized manufacturing jobs have been the foundation of the middle class. The disappearance of well-paid manufacturing jobs in America’s rust belt and elsewhere – owing to globalization and technological change – is at least partly responsible for the rise of authoritarian populism.

Labor productivity in US manufacturing has grown nearly sixfold since 1950, compared to a mere doubling in the rest of the economy. The result has been a striking increase in the manufacturing sector’s ability to produce goods, but also an equally dramatic decline in its capacity to generate jobs. While value added in manufacturing (at constant prices) has broadly kept pace with the rest of the US economy, six million manufacturing jobs have been lost since 1980, while 73 million non-farm jobs have been created elsewhere (mainly in services).

When Donald Trump took office in January 2017, the share of US manufacturing in non-farm employment was 8.6%. When he left office, that figure had fallen to 8.4%, despite his attempt to shore up employment through import tariffs. And despite Biden’s significantly more ambitious efforts, manufacturing employment has dropped further, to 8.2%. The decline in manufacturing employment as a share of total employment (even if not in absolute terms) seems to be an irreversible trend.

A skeptic might object that Biden’s policies have not fully borne fruit and are not yet captured in official statistics. But the fact is that hugely capital-intensive semiconductor plants generate few jobs, relative to the physical investment they require. TSMC’s three fab investments in Arizona are expected to employ a mere 6,000 workers – which works out to more than $10 million per job. Even if the projected tens of thousands of additional jobs in supplier industries materialize, that is a paltry return for employment.

Moreover, one looks in vain around the world for successful examples of reversing the de-industrialization of employment. Germany has a larger manufacturing sector than the US, relative to the size of its economy, but the share of manufacturing employees has dropped like a rock. South Korea has achieved the remarkable feat of steadily increasing manufacturing’s weight in the economy in recent decades, but this has not prevented the sector’s share of employment from declining. Even in China, the world’s manufacturing powerhouse, employment in the sector has been falling for more than a decade, both in absolute terms and as a share of total employment.

It is difficult to avoid the conclusion that boosting manufacturing employment is like chasing a fast-receding target. The world has moved on, and the nature of manufacturing technologies has changed irrevocably. Automation and skill-biased technology have made it extremely unlikely that manufacturing can become the labor-absorbing activity that it once was. Whether we like it or not, services such as retail, care work, and other personal services will remain the primary engine of job creation. That means we need different types of good-jobs policies, with a greater focus on fostering productivity and labor-friendly innovation for services.

This is not to suggest that the CHIPS Act or other policies to boost manufacturing are necessarily misplaced or flawed. They may well strengthen the country’s manufacturing base and promote greater innovation. But rebuilding the middle class, generating enough good jobs, and reinvigorating declining regions call for an entirely different set of policies.

Dani Rodrik, Professor of International Political Economy at Harvard Kennedy School, is President of the International Economic Association and the author of Straight Talk on Trade: Ideas for a Sane World Economy (Princeton University Press, 2017).

© Project Syndicate 1995–2024