NEW YORK ― Citizens in the world’s richest countries have come to think of their economies as being based on innovation. But innovation has been part of the developed world’s economy for more than two centuries. Indeed, for thousands of years, until the Industrial Revolution, incomes stagnated. Then per capita income soared, increasing year after year, interrupted only by the occasional effects of cyclical fluctuations.

The Nobel laureate economist Robert Solow noted some 60 years ago that rising incomes should largely be attributed not to capital accumulation, but to technological progress ― to learning how to do things better. While some of the productivity increase reflects the impact of dramatic discoveries, much of it has been due to small, incremental changes. And, if that is the case, it makes sense to focus attention on how societies learn, and what can be done to promote learning ― including learning how to learn.

A century ago, the economist and political scientist Joseph Schumpeter argued that the central virtue of a market economy was its capacity to innovate. He contended that economists’ traditional focus on competitive markets was misplaced; what mattered was competition for the market, not competition in the market. Competition for the market drove innovation. A succession of monopolists would lead, in this view, to higher standards of living in the long run.

Schumpeter’s conclusions have not gone unchallenged. Monopolists and dominant firms, like Microsoft, can actually suppress innovation. Unless checked by anti-trust authorities, they can engage in anti-competitive behavior that reinforces their monopoly power.

Moreover, markets may not be efficient in either the level or direction of investments in research and learning. Private incentives are not well aligned with social returns: firms can gain from innovations that increase their market power, enable them to circumvent regulations, or channel rents that would otherwise accrue to others.

But one of Schumpeter’s fundamental insights has held up well: Conventional policies focusing on short-run efficiency may not be desirable, once one takes a long-run innovation/learning perspective. This is especially true for developing countries and emerging markets.

Industrial policies ― in which governments intervene in the allocation of resources among sectors or favor some technologies over others ― can help “infant economies” learn. Learning may be more marked in some sectors (such as industrial manufacturing) than in others, and the benefits of that learning, including the institutional development required for success, may spill over to other economic activities.

Such policies, when adopted, have been frequent targets of criticism. Government, it is often said, should not be engaged in picking winners. The market is far better in making such judgments.

But the evidence on that is not as compelling as free-market advocates claim. America’s private sector was notoriously bad in allocating capital and managing risk in the years before the global financial crisis, while studies show that average returns to the economy from government research projects are actually higher than those from private-sector projects ― especially because the government invests more heavily in important basic research. One only needs to think of the social benefits traceable to the research that led to the development of the Internet or the discovery of DNA.

But, putting such successes aside, the point of industrial policy is not to pick winners at all. Rather, successful industrial policies identify sources of positive externalities ― sectors where learning might generate benefits elsewhere in the economy.

Viewing economic policies through the lens of learning provides a different perspective on many issues. The great economist Kenneth Arrow emphasized the importance of learning by doing. The only way to learn what is required for industrial growth, for example, is to have industry. And that may require either ensuring that one’s exchange rate is competitive or that certain industries have privileged access to credit ― as a number of East Asian countries did as part of their remarkably successful development strategies.

There is a compelling infant economy argument for industrial protection. Moreover, financial-market liberalization may undermine countries’ ability to learn another set of skills that are essential for development: how to allocate resources and manage risk.

Likewise, intellectual property, if not designed properly, can be a two-edged sword when viewed from a learning perspective. While it may enhance incentives to invest in research, it may also enhance incentives for secrecy ― impeding the flow of knowledge that is essential to learning while encouraging firms to maximize what they draw from the pool of collective knowledge and to minimize what they contribute. In this scenario, the pace of innovation is actually reduced.

More broadly, many of the policies (especially those associated with the neoliberal “Washington Consensus”) foisted on developing countries with the noble objective of promoting the efficiency of resource allocation today actually impede learning, and thus lead to lower standards of living in the long run.

Virtually every government policy, intentionally or not, for better or for worse, has direct and indirect effects on learning. Developing countries where policymakers are cognizant of these effects are more likely to close the knowledge gap that separates them from the more developed countries. Developed countries, meanwhile, have an opportunity to narrow the gap between average and best practices, and to avoid the danger of secular stagnation.

By Joseph E. Stiglitz

Joseph E. Stiglitz, a Nobel laureate in economics, is a professor at Columbia University. His most recent book, coauthored with Bruce Greenwald, is “Creating a Learning Society: A New Approach to Growth, Development, and Social Progress.” ― Ed.

(Project Syndicate)