The case for a limited but important government role
‘Trumponomics” — to the extent that one can discern such a guiding philosophy at this early date — appears to be based on one key goal: growing middle-wage jobs by limiting low-skill immigration, building more infrastructure (construction jobs pay well), and, most important, reviving U.S. manufacturing, in part by radically limiting the transfer of jobs offshore. This helps explain why the president-elect has touted his deal to prevent Carrier from moving about 1,000 manufacturing jobs to Mexico.
The Carrier deal has dominated recent economic news, with responses running the gamut from “This is totally trivial” to “Picking winners only reduces economic welfare.” But the former characterization undervalues the types of jobs in question, and the latter is not always right. Encouraging Carrier to keep jobs can boost overall labor productivity because those jobs are more productive than the average U.S. job. Indeed, the U.S. Department of Labor has found that U.S. wages would have grown more than twice as fast from 2002 to 2007 had the economy not lost good-paying jobs in industries such as manufacturing (a trend most pronounced in swing states that Donald Trump won, including Florida, Ohio, Pennsylvania, and Wisconsin).
That said, efforts to create and retain high-value-added manufacturing jobs in the United States, while important for keeping America economically competitive, won’t be enough to turn around the underlying productivity slump, because more than two-thirds of U.S. output is not traded internationally. What will really make America great is if the Trump administration establishes a comprehensive national productivity strategy.
Labor productivity, a tally of all the goods and services the country produces per hour of work, has been inching up at an anemic rate of just 1.2 percent per year since 2008. That’s half the rate of the prior 13 years. And in the last year, it has fallen. This is why wages and overall GDP growth have stagnated — not, as the Left claims, because the 1 percenters are taking all the gains.
Why worry about productivity and productivity policy? Without faster productivity growth, it will be impossible to raise living standards at the rate Americans expect. Indeed, the United States enjoys one of the world’s highest standards of living because the average U.S. worker today is so productive, producing in one hour what his counterpart a century ago produced in an entire day. If over the next 30 years U.S. productivity grows at the rate at which it grew from 1995 to 2004 rather than at the current rate, U.S. GDP (and per capita incomes) would increase by 150 percent instead of the currently projected 43 percent. Higher productivity also would afford the opportunity to enact fiscally responsible tax cuts, as every increase of one-tenth of a percentage point in productivity today adds $50 billion to federal revenues ten years down the line. America’s ability to project power globally is directly related to its ability to increase productivity at a robust rate.
Why do we need to have a comprehensive productivity policy rather than just leaving it up to the market? Won’t the traditional conservative formula of tax cuts, limited regulations, and free trade, maybe coupled with some public investment in basic research and education, do the job?
You would think that since most organizations benefit from increasing their productivity, market forces should provide plenty of motivation for them to do so. But in fact a host of market failures have held down productivity growth, making it necessary to adopt a national productivity strategy.
The first failure is that a firm does not capture all the benefits of its own investment; other firms, including its competitors, also benefit. As a result, firms make fewer productivity-enhancing investments than is optimal. This is the key rationale for the research-and-development tax credit, which is designed to stimulate additional private R&D spending by increasing the private rate of return from R&D (i.e., what the company earns), bringing it closer to the societal rate of return (what all companies earn from the investment). While economists have long shown that firms underinvest in R&D absent government incentives, new research suggests that companies capture only about half of the total benefits that come from their investments in new capital equipment. The rest goes to other firms. So firms invest less than we need them to in new, productivity-enabling equipment.
Another market failure relates to uncertainty. Because increasing productivity often depends on adopting technologies that are emerging but not yet fully proven, many potential users will hold off until the payoff is clear. Economists refer to this challenge as “excess inertia” or, more commonly, the “penguin effect” — because, in a group of hungry penguins, most are loath to be the first to test the water, lest they become dinner for a predator. But if no penguin dives in, the whole group risks starvation.
A host of market failures have held down productivity growth, making it necessary to adopt a national productivity strategy.
A third market failure results from U.S. capital markets’ having (d)evolved to reward short-term investing, leading firms to skimp on longer-term productivity-enhancing investments. Business R&D funding has shifted from basic and applied research to less risky and less productive product development, and publicly traded firms invest less in machinery and equipment than do privately owned ones.
A further challenge is that not all industries have the same incentive to raise their productivity. One reason is the “principal-agent” problem: Productivity increases can hurt those who would implement them. For example, while Ford and Toyota managers have strong incentives to adopt production automation, they might be slow to adopt managerial automation. In industries where workers “control the means of production,” particularly professionalized ones — e.g., legal services, accounting, health care, real estate, optometry, and higher education — the workers negatively affected by automation are often the same ones making decisions about whether to adopt it. In these cases, increased productivity often means that their own jobs are eliminated. Why, for example, would law firms not oppose competitors’ provision of more-efficient legal services online, since it would result in fewer lawyers’ practicing law?
Bear in mind that for-profit firms account for no more than two-thirds of U.S. output and that even some of these, such as those in the health-care industry, are heavily influenced by government policy, for better or worse. So any productivity strategy must include ways of improving productivity in government organizations, non-profit organizations, and for-profit industries that are deeply influenced by government. In particular, there is opportunity to boost productivity at all levels of government, which now uses information technology more to improve the quality of services than to cut the payroll — it often includes many unionized workers who resist IT-based automation.
Finally, many of the most important technologies driving productivity involve system interdependence. The effect is that organizations won’t implement a technology unless others have already made associated technological changes. For example, appliance manufacturers might make a “smart” appliance that turns on only when the cost of electricity is below some threshold, but why would they if the U.S. electric grid is not yet capable of time-of-day pricing? Why would people want personalized digital health records when little of the information doctors have is digitized? This chicken-or-egg dynamic exists for many other technologies, such as mobile payments, “intelligent” transportation systems, digital signatures and electronic IDs, and data-driven artificial intelligence. This is why, as described below, the Trump administration needs to put forward digital-transformation policies for these kinds of technologies. This is where smart government policies should come in.
To meet the productivity crisis, it’s important first to understand its cause. Why has U.S. productivity growth stagnated? Most economists are little help here. Liberal economist Alan Blinder voices the thoughts of many in the field when he writes that “what’s scary” about the slowdown “is that we don’t know why” it’s happening. Similarly, conservative economist Greg Mankiw writes that “policy makers don’t have a lot of policy measures they can use to change pretax incomes” — in other words, to grow productivity.
What they are really saying, to paraphrase economist Moses Abramowitz, is that productivity is “the measure of their ignorance.” Others, however, especially if their focus is at the meso-economic level of industries and firms, where most of the action related to productivity occurs, have at least an inkling of the slowdown’s causes. “Most theories of growth are developed at the macroeconomic level — at 30,000 feet,” Clay M. Christensen and Derek van Bever write in the Harvard Business Review. “That perspective is good for spotting correlations between innovation and growth. To understand what causes growth, however, you have to crawl inside companies — and inside the minds of the people who invest in and manage them.”
In other words, you have to look in detail at firms, industries, and technologies. Take quality-adjusted investment in business equipment. It’s about 30 percent less as a share of GDP than it was three decades ago. Some of this decline may reflect a slowdown in the pace of tool improvement, because, outside the Internet and cellphones, the pace of technological innovation appears to have slowed, and much of current innovation is not focused on automating work, which is a key driver of productivity improvement. This trend has compounded the chicken-or-egg dynamic.
So what to do? If the Trump administration’s economic policy is focused predominantly on trade and restoring U.S. manufacturing, that could improve the country’s economic competitiveness, but it won’t be enough to get us out of our productivity crisis. It will happen only when we figure out how to produce more with less in most or all sectors of the economy, and it will happen faster if government policy is well organized to encourage it.
The first step is to get basic market conditions right: Maintain a strong rule of law, respect for property rights, reasonably competitive markets (too little competition limits productivity, but so too can too much), and limits on economic regulation. To that list, we should add a policy principle of firm-size agnosticism. Policymakers have long romanticized and distinctly favored small firms, which, in a system that could fairly be called “small-business cronyism,” enjoy preferential treatment in regulations, the tax code, spending (e.g., subsidized Small Business Administration loans), and procurement (the array of small-business set-asides whenever the federal government wants to buy something). Such policies lower U.S. productivity because, on average, small firms are far less productive than large firms.
The Trump administration could also revive a cultural attitude shared by most Americans until the last few decades: Embrace technological innovation and productivity. Seek to make the proverbial “better mousetrap.” Too many elites have begun to view innovation with suspicion, decrying it for, among other reasons, causing job loss.
Too many elites have begun to view innovation with suspicion, decrying it for, among other reasons, causing job loss.
On these points, most conservatives will probably agree. But more is needed, some of it heterodox from a conventional conservative perspective. Boosting productivity will require more and better “factor inputs” — things that a firm relies on to become more productive, including infrastructure, STEM education, and scientific and engineering research. With some of these inputs, such as education and infrastructure, the question of whether government provides what is needed directly (e.g., through public schools and government-owned roads) or provides incentives for its private production (e.g., through school vouchers and private toll roads) is tactical, not strategic. In other words, conservatives can have arguments about how government policy should be structured, but most conservatives should agree on the need for policy. For example, one can make a compelling argument that school vouchers and more private infrastructure would result in better outcomes, but it’s hard to make a good argument that better schools and infrastructure are not needed. In other cases — most notably, basic and applied research — only increased government funding can fill the gap.
But spurring the expansion and improvement of factor inputs also won’t be enough in itself to maximize productivity growth. The administration and Congress also need to promote an array of policies focused explicitly on productivity. First among them should be policies designed to encourage firms to invest more in machinery, equipment, and software to replace labor.
Wait, policymakers will say. What about jobs? Won’t automation eventually leave most everyone unemployed? No, just the opposite. History, logic, and scores of economic studies all show that higher productivity is associated with faster job growth, not slower. Faster productivity growth frees up purchasing power, which in turn creates jobs. It stirs “animal spirits,” as Keynes called them, and encourages companies to invest and people to spend, creating a virtuous cycle of yet more economic activity and job creation. There is no reason to believe this will change in the future, despite what some techno-utopians claim about the transformative power of technologies such as artificial intelligence and robots. This dynamic of growth leading to purchasing power leading to jobs will exist as long as human needs remain unmet. And given that U.S. median income could increase by a factor of 10 (to $510,000), with most people still having plenty of things to spend that money on (e.g., better vacations, a new boat, eating out more, a college education for their kids), we have a long way to go before we have to worry about running out of work (and even then productivity gains would be used to shorten the work week).
Anything that lowers the ratio of capital cost to labor cost will increase the substitution of technology for labor. So it would be a mistake if, in the forthcoming effort to reform the corporate tax code, existing but weak incentives to invest in capital equipment are jettisoned to offset the budgetary costs of a lower statutory rate. The Trump administration should instead seek to expand investment incentives by letting companies expense for tax purposes all the costs of capital equipment in the first year of use — or, if it is feeling particularly bold, by reinstating the investment tax credit that was eliminated in the 1986 tax reform.
Another way to lower the capital-to-labor cost ratio is to raise the cost of labor. One way to do that is to raise the minimum wage, albeit modestly. The former CEO of McDonald’s reported that, in response to the possibility that the minimum wage will be raised to $15 an hour, McDonald’s began accelerating its deployment of self-serve kiosks and other automation technologies. Wonderful! Fewer low-wage jobs. Another way to lower the ratio is to limit low-skill immigration, as Trump has committed to do. Economists Orachos Napasintuwong and Robert D. Emerson estimate that limiting low-wage immigration would stimulate the adoption of labor-saving technology, with an increased substitution of capital for labor. If it’s harder to hire immigrant construction workers, a firm might be more likely to buy a robotic bricklayer. If California tomato farmers have a harder time hiring immigrant farm laborers, they might be more inclined to lease an automated tomato harvester.
Next, we need to accelerate the pace of innovation in automation technology. The possibilities are endless: robots that could replace janitors, waiters, and housekeepers; artificial-intelligence systems that could replace accountants and insurance agents; autonomous vehicles to replace car and truck drivers; new materials, such as longer-lasting house paint, to reduce the need for maintenance and repair workers; and better drugs to cure diseases, reduce health-care costs, and reduce the demand for medical workers.
As a share of GDP, federal support for scientific research is lower than it was before Sputnik. Moreover, today’s researchers shy away from R&D that could increase productivity, fearing political backlash. The National Robotics Initiative, run by the National Science Foundation, is focused only on robotic technology that complements workers, and not on technology that replaces them. After determining the areas of scientific research most likely to raise productivity, the administration should significantly increase its funding to them.
In addition to facilitating the development of better tools, a national productivity policy needs to promote their wider use, particularly in the case of those, such as drones and an electrical grid with time-of-day pricing, that have myriad applications across the economy. Government officials should be required to take productivity impacts into account when promulgating regulations (or reducing them). The federal government should be a lead adopter of the relevant technologies, using them to accomplish its own mission more efficiently. For example, many government agencies still have paper (or PDF) forms that citizens must fill out and fax or mail in. Simply catching up with the private sector and requiring that all forms be Web-enabled would boost efficiency.
A national productivity strategy must reflect how opportunities and constraints differ by industry. The government needs to understand better why productivity has fallen in, for example, the construction industry over the past two decades. The administration could help reverse the trend by giving construction companies stronger incentives to use building-information-modeling systems and other advanced technologies when erecting federal buildings and infrastructure. The administration also could support industry-led R&D consortia, such as the currently underfunded consortium Fiatech, which works on technology projects to boost construction-industry productivity. Finally, the administration should provide incentives for state and local governments to collaborate to streamline and align their varying building codes to promote industry productivity.
Construction is not the only industry with structural flaws that limit productivity improvement. Health care, transportation, higher education, financial services, and of course government itself could benefit from productivity strategies. Sometimes the federal government will need to spur disruption of recalcitrant industries. U.S. higher education is ripe for it in the form of massive open online courses, but such courses won’t be widely instituted without government pressure, because no university wants to cannibalize its traditional courses and faculty. If the federal government began accepting, as qualifications for employment, alternatives to traditional college degrees, separating the instruction function of higher education from its credentialing function, a robust private-sector credentialing market might develop, driving significant productivity improvement in higher education.
Simply identifying good productivity policies is not the same thing as building the institutional competence and political will to develop and implement them. Concern with productivity is largely absent from the missions of all federal economic-policy institutions, which focus more on managing the business cycle. At present, boosting productivity should rank as more important, and the administration should make doing so the principal goal of its economic policy. It should establish a productivity commission, staffed not by economists but by engineers and business-administration experts, to analyze and advocate productivity policies and programs. It should also require every cabinet agency to develop a plan, within six months, for actions that will help drive productivity growth.
The recent economic slowdown has led pessimistic purveyors of the “dismal science” to tell us to get used to it. Larry Summers says we are in an age of structural stagnation. Robert Gordon laments the “fall of American growth.” James Bullard, president of the Federal Reserve Bank of St. Louis, says that slow growth is the “new normal.” But President-elect Trump didn’t win by listening to conventional wisdom — and if he really wants to create good middle-class jobs, he shouldn’t start listening to it now. Instead, he should embrace a meso-level national productivity policy focused on firms and industries.
It will be much easier for the president-elect to do this if conservatives abandon their blind devotion to laissez-faire economics and put pragmatic productivity policies front and center. That would offer political opportunities as well, because the Democrats have largely given up on productivity growth, seeing it as something that helps only the 1 percent. By seizing this ground, Republicans can reinforce their standing as the party of growth. Conservatives should be forewarned, however, that they have to decide what’s more important to them: untrammeled freedom or long-term prosperity, for some of these steps will require more government involvement in the economy, albeit to help business and spur growth. As the third great wave of industrialization and concentration was beginning to transform the U.S. economy in the 1930s and ’40s, Friedrich Hayek wrote, with respect to the growth of corporations and government, “Personally, I should much prefer to have to put up with some such inefficiency than have organized monopoly control my ways of life.” In other words, Hayek preferred small government and inefficiency to more-active government and more productivity growth. Trump voters seem to prefer the reverse: They want to make America great by restoring robust economic growth, even if it means that government plays a more active role in some sectors. Conservatives should respond by taking the best of free-market economics and combining it with a pro-productivity pragmatism.
— Robert D. Atkinson is the president of the Information Technology and Innovation Foundation.