Principles for US Industrial Policy

In “Design Principles for American Industrial Policy,” (Issues, Spring 2021), Andrew Schrank calls for new design principles with which to anchor new innovation and industrial policies. To have a sustained positive impact, he notes, those policies must create a wide coalition of actors supporting them. This is a truly important insight, and Schrank has demonstrated it across an array of policies over several decades. It is clear that the United States will need to heed those lessons and build new policies along the lines of the targeted-universalism design principles he favors.

Where I would add to Schrank’s contribution is by focusing on the sociopolitical ideals with which we should employ those design principles. After 50 years of growing inequality and decreasing social mobility, the United States has a dual window of political opportunity. The reality is that the majority of Americans now face significant economic insecurity and fear for their future and the future of their children. Hence, Americans want a stronger, but also fairer, nation where everyone has a real shot at the American dream.

For that reason I argue that the United States employ distribution-sensitive innovation policies (DSIPs) as its sociopolitical design principle. DSIPs are designed to reach the dual goals of increasing economic growth while enhancing economic distribution. Amos Zehavi and I have examined such policies in multiple countries of the Organisation for Economic Co-operation and Development, and our findings dovetail with Schrank’s insights. DSIPs can be successful, but their survival depends on crafting a political logic that addresses current political needs and creates a constituency that welcomes their efforts, becoming politically mobilized to ensure their survival.

In his article Schrank mentions two modes of DSIPs, those that aims at low-skilled manufacturing workers and those aimed at the economic periphery. He showed how such programs—for example, the Manufacturing Extension Partnerships, based at the National Institute of Standards and Technology—have been achieved their policy goals, but they have done so only by creating and mobilizing a political coalition to support them. Let me offer two other domains of DSIPs to consider.

Minorities. Governments intent on better integrating members of disadvantaged minorities into the workforce tend to focus on the low end of labor markets. But real progress happens when minorities get into the growing and innovative sectors of the economy. It is not enough to get disadvantage minorities into STEM education; it is also necessary to get them into innovative activities in technology-intensive workplaces. Minority group pioneers can play a critical role by serving as role models in their communities and by becoming nodes in social-professional networks that help future generations navigate the world of technology-intensive industries. Further, the success of such programs creates its own newly empowered political supporters.

People with Disabilities (PWDs). With the United States’ rapidly aging population, the percentage of PWDs is constantly rising. Alarmingly, labor market participation rates for PWDs are very low. More than ever, new technologies hold the promise of enabling PWDs better incorporation into the workforce. Governments can help by pushing for their development and implementation. As the political battles around Medicare demonstrated, older people comprise one of the nation’s strongest political forces, and more and more of them are becoming PWDs.

Schrank has powerfully demonstrated the need to apply targeted-universalism as the core design principle. At the same time, it will be important to ensure that more and more people can actively participate in the economy and fulfill the American dream.

Dan Breznitz

Andrew Schrank compares two ill-fated federal industrial policy programs to three others that are still alive and kicking. He compellingly argues that the difference between the failure of the former and the success of the latter was not in their economic effectiveness, but in their political viability. Contrary to the common wisdom that programs that evade attention are the most resilient, Schrank argues that the path to political viability depends on the different programs’ ability to foster broad constituencies.

Building broad constituencies requires federal programs to adopt a “targeting within universalism” design in which universalism guarantees that all relevant program clients get something and the economically least-developed are targeted to receive a disproportionately higher share of funding than others. However, unlike in social policy programs, targeting in industrial policy is not required to further essential program goals, but to acquire the support of actors—often the representatives of economically weaker states—that would hardly benefit from program allocations rewarded according to purely competitive criteria.

While I readily agree with Schrank that universalism is required to build broad support for programs, I wonder whether targeting is necessary from a political perspective. It is likely that as long as a state receives an equal share of program funding, it would extend its support for the program. Hence, targeting—that is, allocating outsized shares to the less-developed—is unnecessary, at least from a political standpoint.

However, as Schrank duly notes, industrial policy is not exclusively about promoting industry competitiveness. In an era of rising inequality in general, and rising spatial inequality more specifically, governments are seeking ways to narrow the gaps and jump start economic development in “left behind” regions and towns. While it is true that return on government investment tends to be higher in economic powerhouses such as California, from a social equity perspective investing in less-developed Arkansas is the higher priority. I would argue therefore that the rationale for targeting (within universalism) is primarily furthering social goals. All states should benefit from funding to create a broad constituency; targeting is required to address growing social inequities.

Regardless, Schrank’s broader message that for industrial programs to succeed they must expand their constituencies is apt. Indeed, following this reasoning, for industrial policy programs to gain and retain political viability they should be designed to be inclusive. For instance, engaging unions in these programs—as is done, for example, in Germany—could bring a significant new constituency into the fold.

Of course, doing this, or more generally initiating new programs, is no mean task in today’s politically polarized age. Nevertheless, the Senate’s recent passage of the $250 billion US Innovation and Competition Act offers hope that given economic challenges (think China) and widespread social plight, industrial policy is on the rise again. Schrank offers sound advice about program design principles that if followed would increase the likelihood that these new programs would survive and thrive in the coming decades.

Amos Zehavi

Andrew Schrank makes a series of excellent points about the contemporary industrial policy discussion in the United States. I have considerable sympathy for what he says and regard the worries that he addresses concerning the ability of the US political system to design an effective set of policies benefitting American businesses and their workers to be of central contemporary political importance. I offer two thoughts in reaction to his argument.

First, how uncompetitive are American companies that are still in business? The US decline in manufacturing employment has to do with the emergence of China and with secular improvements in productivity, more with the former than with the latter. The United States is still one of the largest manufacturers in the world. Loss of industries, firms, and employment due to the China shock has left us with a surviving manufacturing sector that is relatively lean, and fairly competitive. It is just that the surviving relatively competitive manufacturing sector that we have does not generate a great deal of employment. It would be important to know if Schrank wants policies that will make these already competitive companies even more competitive, which would benefit existing companies, but likely have only modest employment effects. Or if he wants to create more companies so that more people will be employed in manufacturing. Industrial policy will address the first problem, but it’s not clear that it is the right tool for the second problem.

Second, if multiplier effects from manufacturing on the generation of jobs are a key reason to be concerned with the health of manufacturing, wouldn’t it be important to generate an industrial policy that paid explicit attention to the ways in which employment and competitiveness are entangled across manufacturing, service, financial, and even agricultural domains? In some ways, Schrank’s premise—the nation needs a politically feasible but perhaps not so efficient industrial policy—is undermined by his focus on manufacturing alone. Industrial policy is a targeted program, not a universal one. By his own analysis, this is likely to generate opposition from those domains that are not targeted. Schrank wants those drawing up industrial policy plans to take the tension between universality and particular benefits into account, but there are a variety of universals and particulars in play. Is he focusing on the right ones?

Gary Herrigel