Industrial policy, AI and growth models

Michael Spence Says More…

Project Syndicate (PS): You, Anu Madgavkar, and Sven Smit recently pointed out that “economic dynamism and improvements in living standards are vital both to finance climate action and to ensure adequate public support for it”. In your new book,” Permacrisis: A Plan to Fix a Fractured World”, co-authored with Gordon Brown and Mohamed A. ElErian (with Reid Lidow), you highlight major growth headwinds, including “trends that have reduced the supply elasticity of the global system”. In what ways should this new supply environment change how we think about economic growth and stability?

Michael Spence (MS): The last two decades brought a massive increase in productive capacity, as rapidly growing emerging economies, especially China, were integrated into the global economy. As a result, the supply side was not a significant constraint on growth. In fact, global growth remained largely robust even as productivity declined, though there were, of course, some setbacks, such as during the 2008 global financial crisis.

This has changed. Emerging-economy growth is a less powerful deflationary force, and it continues to fade. Whereas the global population used to be relatively young, major economies, including China, together accounting for over 75 per cent of world economic output, are now aging rapidly. The post-1945 Baby Boom generation has reached retirement age and is leaving the workforce, but continues to consume. Labour-market behaviour has changed, with many workers avoiding jobs that are inflexible, stressful, dangerous and often low paid. Labor shortages are affecting all the major employment sectors.

Moreover, global supply networks are being transformed, as countries respond to a long series of shocks — from climate-related weather events and the COVID-19 pandemic to wars and geopolitical tensions — by shifting their focus from efficiency and comparative advantage to resilience and security. And of course, productivity continues to stagnate.

As a result of these secular trends, the global economy is supply-constrained for the first time in 2-3 decades. Growth now depends on relaxing these constraints, such as by finding new sources of labour and productive capacity or by using powerful new technologies like artificial intelligence to boost productivity substantially.

At a macroeconomic level, mindsets and policies also need to shift. For the first time since the early 1980s, we are living in a significantly inflationary environment. When the latest bout of inflation began in mid-2021, central banks largely blamed transitory supply-side blockages from the pandemic, which caused shortages of crucial inputs amid regular lockdowns implemented by China, the world’s leading exporter. But even as those blockages were cleared, the secular trends constraining supply remained, so the major central banks were forced, after a late start, to implement the sharpest and fastest series of interest-rate hikes of the last 30 years.

As interest rates rise, fiscal space is narrowing. This is something governments can hardly afford, given that the fiscal response to the pandemic — like the response to the 2008 global financial crisis — brought a sharp increase in sovereign-debt levels globally. Global sovereign debt now exceeds 100 per cent of global GDP, but investment in major supply-side changes is nonetheless vital to deliver both productivity growth (through structural and technological change) and sustainability (by accelerating the clean-energy transition).

In short, in today’s world, macroeconomic management and fiscal and monetary policy must focus far more on supply conditions.

PS: In Permacrisis, you describe the new growth model that the world needs to “restore productivity, enhance growth potential and achieve our global goals”. What are the pillars of this new model, and what challenges must it overcome to succeed in developing economies with weaker institutions?

MS: Notwithstanding the supply-side headwinds, there are scientific and technological innovations that, if properly used and widely deployed, can produce a surge in productivity. They can also support progress on other dimensions of well-being, such as in healthcare and education. By expanding access to essential services, moreover, such technologies would bolster inclusion, all while supporting the transition to sustainable growth patterns.

Permacrisis devotes considerable attention to these technologies, why they are increasingly accessible and affordable, and how they can transform the supply side of the global economy. They can be found not only in the digital domain, but also in the less-noticed revolution in biomedical and life sciences and in the energy transition. For example, generative AI, the culmination of a stunning sequence of breakthroughs, from language comprehension to image recognition — creates the possibility of equipping people and systems with powerful digital assistants and collaborators across the economy. Of course, tapping the full potential of such technologies will take time, and require major changes in jobs, skill, business models, and organisational structure. We are now in a critical period of intense exploration and experimentation.

A second important ingredient in the new growth model, especially for developing countries, is the maintenance of a relatively open global economy and support for the multilateral institutions that play key governance and financing roles. This does not mean restoring the “old” post-war model, which would be impossible: shocks, national-security concerns, strategic competition between the US and China, and the trend toward diversification cannot be wished away. But we must avoid a relatively comprehensive form of deglobalisation, let alone total fragmentation, outcomes that would hamper developing-economy and global growth and render the transition to a sustainable economic model impossible.

Finally, low-income countries need help. Growth is hard enough in a benign global environment. Under today’s conditions — including shrinking fiscal capacity, rising interest rates, precarious financial conditions, fragmentation of cross-border flows of capital, goods, people  and ideas, and increasingly frequent and severe climate shocks, it is far harder. Given that low-income countries did not create any of these hurdles, they need and deserve greater support from advanced and middle-income economies.

PS: Industrial policy, you argue, “can be essential to a country’s long-term economic survival” and has a role to play in the new growth models, especially to “surmount the infrastructure and innovation challenges we face”. Which lessons from successful industrial policies should policymakers internalise as they seek to ensure that productivity growth is a “central focus” in the development and application of artificial intelligence in the coming years?

MS: Industrial policy has a checkered history. There are notable successes: the diffusion of hybrid corn in the United States, the “green revolution” in Indian agriculture, and the use of foreign direct investment to jump-start the export-led growth model in a number of Asian economies. But there are also failures: mistargeted public-sector investment, misallocation of resources owing to pressure from special interests, extreme import-substitution models, and expansive government intervention in areas where markets and private-sector initiative work better.

These shortcomings point to an ideological component in discussions of industrial policy, related the proper role of government in the economy. But I find this largely unhelpful. Governments have an essential role to play, and it is constantly evolving. The effectiveness of interventions depends accurate identification of the problems, and on wise design, attention to incentives, and effective implementation. And in some areas, interventions are essential.

Consider AI. As James Manyika and I argue in a forthcoming article, AI can produce a huge surge in productivity across the entire economy, but ensuring that it does will require us to avoid the bias towards automation (what Erik Brynjolfsson calls the Turing Trap). This is a powerful bias. Since we often measure the effectiveness of AI applications according to human performance, it would be natural to assume that once AI outperforms humans on average, it should be used to replace humans. Any mention of generative AI is typically met with the same reaction: it is coming for our jobs.

This bias is a mistake. The best and most natural use of a range of AI technologies, including generative AI, is in the form of augmentation, or what might be called the “powerful digital assistant model”. But avoiding the automation bias, ensuring that workers have the skills they need, and communicating effectively with employees and other stakeholders will require more than just technological prowess.

Likewise, the diffusion of AI across the economy is not assured. Previous rounds of digital adoption were characterised by a pattern of divergence, with some sectors (like tech and parts of finance) moving rapidly, and others lagging. If the same pattern emerges with AI, for example, large innovative firms embrace AI applications and use cases, leaving behind smaller businesses with fewer resources, the economy-wide productivity surge that is needed will not materialise. Policy, business (especially the tech industry), and the education sector all have a role to play in ensuring widespread accessibility and diffusion.

As it stands, the policy agenda is heavily weighted toward downside risks, potential misuse, biases, and the rights surrounding the use of data, including “public” data, on the internet. These are important and very real issues. But so is the need for training, skills, and access. The policy agenda should reflect this, with a better balance between avoiding negative outcomes and promoting positive ones.

By the way…

PS: To “unleash the potential” of new growth models, you explain, we also need a “new model for economic governance”. This requires that we “rethink the relationship between monetary and fiscal policy”. What are the main shortcomings of the current dynamic, and what might a more constructive relationship look like?

MS: It is mainly about mindsets or what economists might call “implicit models”. A generation and a half of people in business, finance, and policymaking have lived entirely in a largely deflationary, demand-constrained world. In the decade after the global financial crisis, and then through the COVID-19 pandemic, we had zero or negative interest rates and huge infusions of liquidity via central-bank asset purchases. Despite occasional warnings, there was no sign of inflationary pressure. Price increases consistently came in below target.

There is a tendency to assume that as soon as the current bout of inflation is reined in, with price growth being stabilised around some target, possibly somewhat higher than the old target, we will return to this world. Inflation will no longer be a threat; real interest rates will be low; and future large injections of liquidity can be pursued without fueling another inflationary surge.

But for the structural, largely supply-side-related reasons described above, this is very unlikely. We (and here I include monetary and fiscal policymakers, as well as the private sector) should thus be preparing for a world with higher costs of capital, lower valuations for a range of assets (including equities, bonds, and real estate), and tighter fiscal conditions, with some economies facing fiscal stress. Moreover, we should be prepared for “accidents” like the US bank failures of last spring. We can expect, however, that the adjustment of policymakers’ mindsets and behaviors to rapidly evolving structural conditions will lag.

PS: In discussing the shift toward “managed globalisation-lite”, you argue that, to avoid a “new Cold War bringing with it profound economic consequences”, we must “distinguish between what’s essential to advance sustained economic growth and what’s necessary to preserve countries’ autonomy”. When it comes to the restrictions imposed on China by the US, have policymakers added entries to either list that do not belong, or left out entries that do?

MS: Here, I think there is room for some optimism. Policymakers on both sides accept that there are national-security considerations that demand restrictions on the flows of strategically important goods and technologies. But there seems to be a growing awareness that these restrictions should be confined to areas with a direct impact on defence and military capabilities. Both seem to want to avoid broad-based restrictions that will have adverse effects on their economies, not to mention the rest of the world, with no evident benefits. Dialogues around these issues have started. This is a positive sign, though ensuring that the details and implementation of measures align with this approach will not be easy.

The fact is that attempting to constrain China’s development across the board is not in anyone’s interest. (Nor is it really possible, though of course some dimensions of China’s development could be slowed down.) A majority of emerging-market countries, now a substantial part of the global economy, have made it clear that they oppose broad-based restrictions on trade and technology, as well as sweeping sanctions. They have not, for example, endorsed the sanctions associated with the Ukraine war.

PS: In “Permacrisis”, you, Brown, and ElErian each brought your own experience and expertise to the table, with minimal “overlap and redundancy”. When reflecting on each of your co-authors’ insights regarding potential solutions or approaches to addressing the permacrisis, which stand out?

MS: The ideas in the book came from extensive discussions, the frequency of which increased during the pandemic (thank you, Zoom!). These are difficult times. No one has complete solutions, or anything close. There are no silver bullets that we can see. But we do know that the challenges ahead are interconnected, so narrow responses will be inadequate.

In fact, while the book is divided into three parts, growth models and changing structural conditions; macroeconomic management; and interdependence, global governance and the global nature of critical challenges like climate change — we do not really view these as separate issues. Instead, they can be viewed as necessary considerations in any effort to ensure broad-based, long-run prosperity. Together, they might even be sufficient.

While we each brought different perspectives and expertise to the table, we all take responsibility for the views expressed throughout Permacrisis. Our hope is that the book will contribute to discussions on how to adapt to new global conditions. We believe that small steps forward on multiple fronts can eventually add up to fundamental change.

As for specific solutions, we highlight several. One is the determined and deliberate use of scientific and technological tools to advance growth, equity, and sustainability. A second is the rebalancing of fiscal and monetary policy to meet the needs of a supply-constrained world (including abandoning the transitory mindset). A third is a revision of the rules, norms, and institutions that underpin global governance. We need reformed and well-capitalised multilateral institutions to organise the global response to global challenges. Rising nationalism and unilateralism risks marginalising them instead. This latter element will probably be the most difficult to implement, not only because of its sheer complexity, but also because in today’s world, no single power calls the shots.

Michael Spence, a Nobel laureate in economics, is emeritus professor of Economics and a former dean of the Graduate School of Business at Stanford University. Copyright: Project Syndicate, 2023. www.project-syndicate.org