Innovation Files has moved! For ITIF's quick takes, quips, and commentary on the latest in tech policy, go to

Two Cheers for Martin Baily’s “U.S. Manufacturing”

Chevy Volt Factory

It was with great interest and mostly pleasure that I read Martin Baily and Barry Bosworth’s new article in the Journal of Economic Perspectives, “U.S. Manufacturing: Understanding Its Past and Its Potential Future.”

The article attempts to analyze recent trends in U.S, manufacturing performance, including output and employment. This is an area ITIF has been working on for a number of years. And in the past, Baily has been skeptical of our analysis, which claimed that U.S. manufacturing was in fact worse off than official statistics, in part because of the overstatement of computers and electronics manufacturing output. So it was with great delight, and some surprise, to see that Baily and Bosworth have now embraced this analysis.  As they note, the fact that measured manufacturing output’s share of GDP has remained stable “is largely due to the spectacular performance of one subsector of manufacturing: computers and  electronics.” In fact, as ITIF showed, they also show that by taking out computers, overall real manufacturing output fell from 2000 to 2011, something that is unprecedented in our almost 250-year history.  They also rightly point out that the massive loss of manufacturing jobs in the 2000s cannot have been due only to superior productivity growth. As they indicate in their article, that productivity growth of non-electronics/computer manufacturing was about on par with overall rates of economy-wide productivity. And they acknowledge that some of the loss of jobs was due to trade and the expansion of the trade deficit, including with China. Given that up until now mainstream economists were largely in denial of this narrative, preferring to believe that all is well with U.S. manufacturing and that the job loss was all due to productivity, this is a major shift and important contribution to the debate.

So if their article is an important contribution to the analysis of manufacturing, their policy prescriptions fare less well. First, they fall back on the standard neo-classical argument that the trade deficit is “largely a macroeconomic phenomena, reflecting the gap between a national savings and domestic investment.”  If they really believe that, then why do they even bother with any of their policy prescriptions, such as lowering the corporate tax rate? By this logic, why not raise the corporate tax rate? After all, their theory would suggest that bad domestic policies could play no role in our trade deficit. According to neoclassical economic trade theory, these factors can have no effect on the ability of business establishments in the United States to thrive in international markets because that is determined solely by our savings rate. By this definition, there is no trade deficit of any size that can provide evidence of competitiveness failure.

But as non-neoclassical economist Robert Blecker states, “This identity does not prove causality, and is consistent with other causal stories about the trade deficit.” In other words, what the conventional story fails to recognize is that savings is a function of national competitiveness. If, for example, the Chinese stopped manipulating their currency, the U.S. trade deficit would fall and the Chinese would buy less of our government debt. The result would be a rise in both U.S. exports and interest rates. And both would spur more savings. Higher interest rates would lead more Americans to save. More exports (and relatively fewer imports) would boost U.S. corporate savings. And more jobs and higher wages through exports (exporting firms pay 9.1 percent more than jobs in firms that export less) would boost individual savings and reduce the budget deficit.

And when it comes to policy prescriptions, they cannot escape their neoclassical roots, which decry “subsidies,” and any economic policy that does not treat all economic activity the same. So they fall back on generalized policies that while helpful, will be insufficient. The first is their call to reduce the budget deficit in the hope that it will boost national savings and reduce the value of the dollar. But we’ve been through this before. In the late 1990s, Congress and the Clinton administration ran budget surpluses, but the value of the dollar went up and the trade deficit skyrocketed. Besides, as noted above, this mechanical link between savings and deficits is largely chimerical.

Though earlier in the paper they scoff at advocates who “blame other countries’ for our manufacturing woes,” they rightly argue that U.S. trade negotiations should do more to pry open foreign markets and limit foreign currency manipulation (I thought currency values were set only by national savings rates. I guess it’s more complicated than this). They rightly call for a cut in the corporate tax rate and shifting to consumption taxes (as ITIF does here). They also rightly argue for better education and skills policies, including vocational education.

But yet they cannot bring themselves to go all the way to supporting the fourth  “T’s” of ITIF’s Tax, Trade, Talent and Tech agenda. They get all the way to step 8 on ITIF’s logic chain to a national competiveness policy but stop there because they believe that all we need are better innovation inputs, like better skilled workers and infrastructure and not a technology strategy.

As we note in that report:

Among those who recognize the need for government to act all along the innovation cycle and in all industries, many believe that this role should be limited to simply supporting “factor conditions” that all firms can benefit from (e.g., free trade, better education, a good regulatory system, basic research, etc.). For them, the problem is not within enterprises; it’s that enterprises lack the necessary inputs for successful innovation. Indeed, most leading reports on innovation and competitiveness reflect this focus. The widely praised Rising Above the Gathering Storm report largely confined itself to arguing for greater government support for factor conditions such as science funding and STEM education. Likewise, a report from the Council on Competitiveness calling for action argued, “Education is perhaps the single biggest threat to future American prosperity.”

In other words, effective innovation policy has to go beyond simply supporting factor conditions that all firms can use; it has to go inside the “black box” of the firm to help firms and key industries thrive. This means that effective innovation policy includes things like much stronger R&D credits; programs to help small and mid-sized manufacturers adopt new technology; policies focused on key technologies (such as advanced batteries, robotics, and artificial intelligence); and policies focused on key industries (like broadband, clean energy, semiconductors, IT, and life sciences) all the while enabling competitive markets. While it is generally inadvisable for governments to pick specific winner companies or narrow technologies, it can and should pick broad technologies and industries to support.

For them there is no market failure on the development of technology, even though economic research has shown that there is. But perhaps this is because they misinterpret what we propose, claiming that while ITIF stresses the need for government support for technology, “U.S. companies remain strong in technology development while viewing it more profitable to produce the technology goods overseas.”  This is a fundamental misunderstanding of the U.S. technology innovation system on two levels. First, many technology companies (e.g. Boeing, GE, Merck, Intel) produce here, since there can be strong reasons to produce advanced technology products in high cost, high skilled regions like the United States.  Government support for product technology development will mean that much of this will be produced here once it is developed. Second, Baily and Bosworth seem to conceive of technology as only product technology.  But process technology is just as important for the future of U.S. manufacturing. In fact, if fully funded by Congress, it is likely that many of the National Network of Manufacturing Innovation institutes would be focused on process technology—the technologies that make it easier and cheaper to produce here.

But regardless of the limits on their policy prescriptions, the fact that Baily and Bosworth make an important analytic contribution on what has really happened to manufacturing—a contribution that hopefully the media, pundits and policy makers will finally pay attention to—it is important for everyone interested in the future of the U.S. economy to read this article.


Print Friendly, PDF & Email