Lecture notes

Day 6

Lectures from Washington, D.C., Monday, August 3, 2026.

Washington, D.C.

Monday, August 3, 2026

3 lectures

Open the full notes (PDF)

Lecture 1

Global Economic Update

A lecture by Indermit Gill

Chief Economist and Senior Vice President for Development Economics, World Bank Group; with Ayhan Kose, Deputy Chief Economist and Director of the Prospects Group, World Bank Group.

The bottom line

Potential growth has fallen for two and a half decades, and a run of shocks has left the world economy expanding below even that diminished potential. The post-pandemic rebound was strong but deeply uneven, convergence has stalled outside China and India, and a record wave of young workers is arriving just as investment and trade weaken. The realistic upside is AI-driven productivity, but capturing it requires trade openness, credible fiscal rules, and a far better business environment.

Growth is slowing, and the shocks keep coming

For a quarter century, growth rates have fallen steadily while the capacity to respond has narrowed under the weight of demographics, rising debt, and trade tensions. The decline in potential growth is rapid: for emerging market and developing economies it has slipped from about 6% a year in the 2000s to 5% in the 2010s and 4% in the 2020s. The world economy has absorbed a great deal and is still standing, but keep growth falling and eventually a single shock could tip it over.

Ayhan Kose framed the era as "one shock after another": the pandemic, supply disruptions, a sharp rise in interest rates, the Russia–Ukraine war, an inflation surge, and now conflict in the Gulf. That conflict is the largest oil-and-gas disruption by some measures in five decades, and its effects cascade, first through fuel, then through food and fertilizer as importers face higher prices and exporters impose bans, and finally through finance as higher inflation lifts policy rates. For food importers the stakes are stark: in Nigeria, the poor spend around 70% of household budgets on food. Because most developing countries are small economies, there is no substitute for more trade, even when larger partners turn inward.

An uneven recovery and a jobs wave

The 2020 downturn was the deepest global recession since 1960 yet drew the strongest recovery, the world economy expanding about 14% over five years. But the rebound was driven by advanced economies. One-third of emerging market and developing economies (and close to 60% of fragile and conflict-affected states) still have lower per-capita income than in 2019. Stripping out China and India, convergence toward advanced-economy incomes has essentially stopped; in Sub-Saharan Africa, per-capita income has been roughly flat for 15 years. When per-capita incomes stop rising, Gill warned, the entire pathology of development changes.

The defining challenge is demographic. Some 1.2 billion young people will reach working age between 2026 and 2035 (the largest youth cohort ever) and the next surge will crest in Sub-Saharan Africa, South Asia, and the Middle East and North Africa, where East Asia's old path of manufacturing and trade is less available. Yet investment and trade growth both run near half their 2000s pace, and investment per worker has stagnated outside China. Growth that does not create jobs breeds political and social strain, so the task is not just more growth, but growth that employs.

Evidence at a glance

Potential growth is falling everywhere. Advanced-economy potential has roughly halved since the 2000s; emerging-market potential has slipped from 6% to 5% to 4%.
2026 is subdued. World growth is expected around 2.5%; emerging and developing economies near 3.6%, about 10% below potential, and only 2.6% excluding China and India.
Debt has surged. Government debt in emerging and developing economies has more than doubled to about 77% of GDP over 10 to 15 years, leaving little room if interest rates rise.
Trade has not collapsed, uncertainty has spiked. World trade rose from about 20% of GDP in 1945 to 60% in 2022; the ratio has since plateaued but not fallen, while trade-policy uncertainty has climbed.
Barriers start at home. Middle-income economies impose higher tariffs and non-tariff barriers on each other than they face in advanced markets, and now account for more than half of world trade.

The policy response

AI is the credible upside, if the fundamentals are in place. In the World Bank's more optimistic scenarios, AI could deliver the fastest productivity growth in decades and reverse the slowdown, but only economies with open trade, sound public finances, and a strong business environment will capture it.

Liberalize trade unilaterally. For a small economy, opening up yields large gains regardless of what the United States or other partners do.
Adopt and enforce fiscal rules. Emerging economies have improved monetary policy but still lag on fiscal management; credible rules raise the odds of high-quality adjustment, as Turkey's earlier turnaround showed.
Fix the business environment. With foreign direct investment near two-decade lows relative to GDP, improving conditions for firms is essential to attracting capital.
Invest in infrastructure and people. World Bank Group's response centers on aggressive investment in physical, digital, and human capital, plus mechanisms to mobilize private capital.
Take AI seriously, especially if poor. AI can lift agricultural productivity and the quality of health, judicial, and legal services; poorer economies are often better prepared than assumed, as small and informal firms already adopt digital tools, and faster growth, anchored in large economies whose spillovers lift others, is what gives the distribution debate anything to work with.

Lecture 2

Industrial Policy for Development: A Global Overview

A lecture by Tristan Reed

Economist, Development Research Group, World Bank Group.

The bottom line

The debate is no longer whether governments should do industrial policy but how. Defining it as government action to grow a strategic business activity, the World Bank's new global report urges countries to match tools to three characteristics – government bandwidth, local market size, and fiscal space – and to reach first for low-cost "public inputs" that address market failures directly before turning to subsidies, tariffs, and other incentives.

Reframing an old debate

The starting point is the 1993 East Asian Miracle report, the World Bank's first policy research report. Written amid a fierce debate over the role of the state versus the market, it credited growth to macroeconomic fundamentals – human capital accumulation, high savings, macro stability, and modest fiscal deficits – rather than to industrial policy, and concluded that even where interventions had succeeded in Japan and Korea they were unlikely to be replicated because, in the words of the foreword, those countries had a unique culture, politics, and history.

This new report reopens the question from a global, rather than East Asian, vantage point, organized around five questions: what industrial policy is, who does it, how to do it, which activities to target, and how to get the institutions right. It defines industrial policy as government action to grow a strategic business activity – where "industry" means any business activity (agribusiness, critical minerals, professional services, tourism), "activity" captures specific tasks and segments of a value chain, and "strategic" signals that government has judged one activity more important than others.

Crucially, developing countries are not bystanders. A survey of the Bank's lead country economists found that 80% had been asked about industrial policy by clients in the past year, overwhelmingly motivated by economic development rather than climate or security. Reading national development plans shows poorer countries list more priority industries – reflecting a hunger for diversification – while tariffs run more than twice as high in low-income countries and subsidy spending is heaviest in upper-middle-income ones.

A toolkit matched to country characteristics

Rather than ask whether to intervene, the report starts with how. It sorts tools into three families: public inputs (industrial parks, skills development, market-access assistance, and quality infrastructure such as standards and metrology), market incentives (subsidies, tariffs, local-content requirements), and macroeconomic interventions (a competitive exchange rate, a general R&D tax credit). Public inputs and direct subsidies are "first choice" because they address market failures directly; tariffs and other indirect tools are "second choice."

Which of these is feasible depends on three country characteristics. Government bandwidth – a workforce able to engage firms and gather information – is needed to tailor public inputs. Local market size, measured by the middle class and any customs union, determines whether economies of scale can be reached behind protection. Fiscal space determines whether cash subsidies are affordable at all. A country short on all three still has options: industrial parks (which can be self-financing), commodity export bans, and exchange-rate devaluation. As bandwidth grows, the cheap first-choice public inputs come within reach; market size unlocks second-choice tools; only fiscal space opens the full menu of subsidies.

Evidence at a glance

Industrial parks can spur local development. A Journal of Development Economics study mapping Africa's parks found, within a 10-kilometer radius, a shift out of agricultural jobs and gains in education and durable goods such as refrigerators, televisions, and mobile phones.
Targeted public inputs work when tied to demand. Costa Rica attracted Intel by co-designing a one-year associate's degree with the firm's HR department and a local university; market-access assistance in Latin America and quality certification (the Benin Republic's pineapples) have lifted exports.
Well-designed subsidies can be both effective and efficient. Korea's heavy-and-chemical-industry credit program raised sales, employment, and – after a decade of learning – total factor productivity, leaving the economy about 3% larger each year (Choi and Levchenko); Romania's software payroll-tax exemption helped make it a leading software-services exporter.
Innovation grants pay off through infusion. Brazil's Embrapa turned the country from a net food importer into an exporter; its Finep grants let firms import and adapt foreign technology, with added payroll-tax revenue covering the cost.
Tariffs rarely succeed alone. The 1890 McKinley tariff (70%) built a U.S. tinplate industry only over a decade, behind a huge domestic market and with domestic inputs – and Irwin argues it would have emerged anyway.
Export bans need market power. Indonesia's nickel ban drew investment in smelting because it held a large global share; an identical bauxite ban failed because it did not.

What to target – and for how long

Target what is new, keep competition alive, and be patient. Because true market failures are hard to measure, the practical rule is to promote activities the economy has not done before – and to expect results over five to ten years, not one political cycle.

Favor new activities. Learning comes from doing something new, yet Kenya subsidizes long-established tea and 92% of Serbia's subsidies go to firms that do no R&D.
Preserve competition. Support dispersed across many firms raises productivity more than picking a single winner (Harrison and co-authors, China); where the domestic market is small, exposure to trade must substitute for domestic rivalry, as Japan's development-era export-disciplined cartels show.
Handle jobs and duration carefully. Broad, easy-to-access adjustment programs (Brazil's reskilling) beat narrow ones (U.S. trade adjustment assistance); the hardest political problem is not letting losers go but letting winners go, as reversals of green industrial policy in the United States and Brazil illustrate.

Lecture 3

Industrial Policy in the Digital Age: Lessons from Asia

A lecture by Siddharth Sharma

Lead Economist, Office of the Chief Economist for South Asia, World Bank Group; with Alessandro Barattieri, Senior Economist, Office of the Chief Economist for East Asia and Pacific, World Bank Group.

The bottom line

Asia is fast-growing but faces a productivity and jobs challenge that industrial policy has so far addressed unevenly. Drawing on the World Bank's April 2026 regional updates for South Asia and East Asia and Pacific, both speakers argue that targeted measures work only atop strong foundations – infrastructure, human capital, and open, undistorted markets – and that artificial intelligence is now reshaping which tools make sense.

South Asia: fast growth, a jobs gap, and an AI test

South Asia grew 7% in FY2025, again the fastest-growing emerging-market region, powered by domestic demand in India and a better-than-expected recovery in Sri Lanka. But it is exposed to conflict in the Middle East: a large net energy importer (energy is about 5% of the consumption basket, above the EMDE average), it also sends about 13% of its exports to the Gulf and hosts roughly 9 million migrants there whose remittances reach about 8% of GDP in Nepal. With crude expected to average $94 a barrel over FY2026, the risk has materialized on the upside.

Against this, 2026 is a year of trade opening. India's new free-trade agreements with the EU and UK cut tariffs on about 95% of traded goods and roughly double the share of its output covered by such agreements to 30%, near China's level; Sri Lanka's phase-out of para-tariffs would cut its average import duty by nine percentage points. First-order estimates show real incomes rising across the distribution – more for rural households, and, in Sri Lanka, more for the poor because manufactured food faces the largest cuts. Yet the jobs challenge is stark: about 16 million young people will enter the labor force each year for 15 years, against roughly 12 million jobs created annually over the past decade.

Artificial intelligence could help or hurt. About 22% of South Asian jobs are highly AI-exposed – below the EMDE average because so many work in agriculture – but those jobs earn about 42% of labor income, and more are complementary to AI (15%) than at risk of automation (7%). Job-posting data confirm it: hiring in highly exposed, low-complementarity occupations fell about 22%, while the export-oriented ICT sector is upgrading toward more AI-complementary skills.

East Asia: a productivity problem and three pillars

East Asia's recent growth outside China has come from capital accumulation, not productivity, and firm-level data show a widening gap between leading firms and the global frontier, especially in digital sectors. That makes the capital-driven, labor-intensive model risky – and makes AI, whose productivity potential is large but uncertain, worth taking seriously.

The East Asia and Pacific update frames industrial policy broadly, as three pillars. Foundational public goods come first: many countries sit below the global median on health and education, and human capital, not algorithms, is the binding constraint. Addressing policy failures – "do no harm before you do good" – means, distinctively for the region, opening highly protected service sectors; Vietnam's post-2007 services liberalization raised labor productivity about 2.9% a year over nine years. Only then come targeted interventions against market failures, from Indonesia's resource downstreaming to Vietnam's plan to train 50,000 semiconductor engineers.

A central distinction runs through the evidence: effectiveness versus efficiency. A policy can hit its target yet cost more than it is worth. Korea's 1970s heavy-industry program was both effective and efficient, with credit flowing to the highest-welfare sectors; China's shipbuilding drive lifted its global market share from 10% to nearly 50% but, by the study's cost-benefit measure, was inefficient. And along the AI value chain, developing East Asian economies sit mostly at the low-value assembly end – a position Korea's own three-act semiconductor history suggests can change only by building capital and technology at home.

Evidence at a glance

Trade openness is rising fast. India's FTAs cover about 95% of traded goods; Sri Lanka's reform would cut its 19% average import duty by nine points. The average real-income gain is about 0.2% in India but about 2% in Sri Lanka, which liberalizes across all partners.
South Asia leans on protection. Industrial-policy measures nearly doubled after COVID; the most common type is trade-related import protection, versus subsidies in the average EMDE, and countries tend to target larger, higher-productivity firms.
Its track record is mixed. Event studies show import-protection measures do cut imports, but export-promotion measures show no significant effect – consistent with limited fiscal space, thin government bandwidth, and missing complementary inputs.
Targeted subsidies need strong foundations. Matching global firm and subsidy data, export promotion and direct transfers correlate with higher productivity only in G20 economies, not in developing ones.

Implications for policymakers

Build the foundations first; target only where market failures are clear. Broad-based reforms – infrastructure, business climate, institutions, open services – do most of the work; targeted tools should complement them, and should always be judged on efficiency, not just effectiveness.

Tilt growth toward jobs with first-best tools. Where failures are well identified, favor industrial parks, skills programs, and market-access assistance – for example, to help the ICT sector adjust to AI-driven reshoring pressure.
Weigh the data-center rush carefully. Because AI compute is internationally traded, location reflects fundamentals – reliable power, water, and, strikingly, governance – and spillovers are uncertain.
Do no harm before doing good. Liberalizing protected services is itself a powerful industrial policy, often safer than new subsidies.