Last year’s US Chips and Science Act created large subsidies for investments in domestic semiconductor fabrication facilities (fabs), on the grounds that microchips are essential both to the US economy and to national security. But while no one disputes the importance of chips (which are used in everything from cruise missiles to refrigerators), there are serious questions about whether subsidizing such investments is the best way to secure a reliable supply.
The failures of Silicon Valley Bank and Signature Bank are significant market events. But, given an overheated labour market and 1970s-like inflation, if the Fed cannot see the whites of the eyes of a systemic banking crisis, then it must move aggressively on the inflation front.
In fact, US competitiveness in chipmaking might deteriorate further because of the legislation. After all, governments do not have a good history of “picking winners”. All too often, such interventions help to prop up losers and inefficient producers, leading to monopolization and concentrations of market power as new (unsubsidized) firms are deterred from entry. Moreover, in the case of chips, the industry has been retrenching with layoffs, cancelled or postponed expansion plans, and other signs of a slowdown.
In response to the new US subsidies, South Korea recently announced plans to support a US$228 billion investment by Samsung to build the world’s largest advanced semiconductor complex; the European Union has taken up a proposal for a €43 billion (US$46 billion) European Chips Act; and other countries have begun to roll out similar forms of support for their own industries. As a result, taxpayers in the US and these other jurisdictions may end up financing a wasteful chip glut.
Nor is a subsidy race the only source of waste. Today’s semiconductor industry is so globally interdependent that almost no chips can be produced without machines and materials from multiple international sources. Achieving self-sufficiency thus would be prohibitively costly even if it was feasible, which it probably is not.
The United States has long been a global leader in chip research and development. It makes many of the machines needed to manufacture advanced chips, and for many years, it accounted for the world’s largest share of chip production. But, over time, it has lost market share to companies in Europe and especially Asia, where production costs are estimated to be 40% lower. Intel, the early pioneer in semiconductors, has publicly acknowledged that it cannot compete without subsidization.
As if inefficient subsidization was not bad enough, US regulators have also announced excessive eligibility criteria for companies applying for the subsidies. Among other things, companies are expected or encouraged to provide childcare for their workers, refrain from stock buybacks, and give forecasts of future profits so that the government can siphon off any excess margins.
All these measures are ill-advised. Requiring companies to pay for childcare cannot fail to make chip production costlier, which defeats the purpose of the subsidy. It makes no sense to single out an industry whose competitiveness is already challenged. Insofar as childcare benefits are attractive to workers, businesses can provide them in lieu of higher pay, or Congress can mandate that all working families have access to it (as it already does with health care).
The stock-buyback provision is also flawed. According to US Secretary of Commerce Gina Raimondo, Chips Act subsidy applicants will receive preferential treatment if they commit not to engage in buybacks for five years. But this will become yet another source of inefficiency.
Until modern times, businesses were mostly family owned, and there were few mechanisms by which a family with a low-return business could invest in another, more profitable one. But a great societal benefit of the modern corporate structure is that it enables investors to shift funds from weak enterprises to those with better prospects.
Once the corporate form was established, some governments taxed both corporate profits and dividend payouts to shareholders. Share buybacks thus became a mechanism by which investors could take their gains and invest in other businesses without incurring double taxation. To discourage buybacks will induce companies to reinvest their profits even if their profit outlook is unpromising. Again, this is not just inefficient but self-defeating. If the goal is to boost American competitiveness, there should not be extra hurdles to allocating investments toward their most promising uses.
Taxing “excess” profits is also likely to have detrimental effects. In tech-centred industries, companies often maintain a diversified portfolio of risky investments in the expectation that while some will be wildly successful, others will fail or struggle. Faced with a penalty for success, chip companies may decide either to locate their risky research overseas or to reduce the risk profile of their research and development.
Companies that avail themselves of the subsidies will surely have to report extensively on their operations to the US government, which in turn will need to monitor recipients and decide whom to offer loans or grants in the first place. Both activities will further increase the costs to taxpayers.
All these unnecessary regulatory provisions should be rescinded. Introducing a chip-production and trading regime among allies would be a far better way to ensure a reliable supply, as would funding programs to train workers in the chip sector. If the point is to secure the supply of a critical input, the last thing the industry needs is the added burden of childcare costs, restraints on payments to shareholders, and excessive taxation.
Anne O. Krueger is senior research professor of international economics at the Johns Hopkins University School of Advanced International Studies, a senior fellow at the Center for International Development at Stanford University and a former World Bank chief economist and first deputy managing director of the International Monetary Fund.
Copyright: Project Syndicate
