Kuznets Lecture New Structural Economics: A Framework for Rethinking Development Justin Yifu Lin Chief Economist and Senior Vice President the World Bank March 1,

Overview of Presentation Why do we need to rethink development The New Structural Economics Industrial Policy and Growth Identification and Facilitation: An application of new structural economics 2

WHY DO WE NEED TO RETHINK DEVELOPMENT 3

Economic Crisis and Crisis in Economics Economic Theory Explain Observed Economic Phenomena Guide Economic Policies or Choices Rethinking Economics Failure to: 4

How has economic development theory evolved? Structuralist Approach Focus on Market Failures: Import Substitution Strategy Miserable results Liberalization Approach Focus on Government Failures: Privatization and Marketization Mixed Results Market based economies with proactive role for government Successful East Asian Tigers: Export Promotion China, Vietnam and Mauritius: Dual-track approach to transition Rethink Development 5

World Bank has been in the process of rethinking economic development Export Orientation and Market Friendly Government (i)Openness; (ii)Macro stability; (iii)High rates of saving & investment; (iv)Market mechanism; (v)Committed, credible & capable government No one-size fits all 6

THE STRUCTURAL ECONOMICS 7

Introducing… New Structural Economics Application of neoclassical economic approach to understand changing economic structure in development Provides a consistent framework for the five stylized facts of Growth Report as well as the findings from the East Asian Miracle and Lessons from the 1990s Contributes to new theoretical and policy insights for economic development 8

Introducing… New Structural Economics Sustainable income growth is the foundation for poverty reduction and development Sustainable income growth is a recent phenomenon 9 The sustainable income growth is a result of continuous technological innovation as well as structural change

Industrial Structure in New England, 1900s 10

Industrial Structure in New England, 1600s 11

Industrial Structure in New England, 1800s 12

New Structural Economics (NSE): Key Concepts The main hypothesis: Industrial structure is endogenous to endowment structure Initial Endowments (determine the economy’s total budgets and relative factor prices at time t) – Comparative advantage – Optimal industrial structure (endogenous). Dynamics: Income growth depends on – Upgrading of endowments – Upgrading industrial structure – “hard” and “soft” infrastructure Following comparative advantage is the best way to upgrade endowment structure and to sustain industrial upgrading, income growth and poverty reduction. 13

New Structural Economics (NSE): Key Concepts (2) Firms maximize profits…choice of technology and industries based on relative factor prices… Need for competitive market system Industrial upgrading needs to – Solve coordination problems – Address externalities Need for a facilitating state 14

NSE and The Growth Commission’s Stylized Facts Policy Recommendation from NSE – Following comparative advantage : Conditions Market economy Facilitating State The results: – Openness and advantage of backwardness – Competitiveness and strong external as well as fiscal accounts: fewer home-grown crises and larger scope for countercyclical fiscal policies. – Large economic surplus and high returns to investment: high rate of savings and investment. The NSE’s recommendations are consistent with the East Asian Miracle’s findings. Growth Report Stylized Facts: #4 #5 15 #1 #2 #3

“No one size fits all” then “What size fits what?” New theoretical insights from the NSE: New structural economics emphasizes that countries at different levels of development have different optimal industrial structures, firm sizes, capital requirements and nature of risks, therefore, many institutions and policies should be different accordingly and have different policy insights compared to the old structuralism and neoclassics: Financial institutions: – Old structuralism: Direct government mobilization and allocation of financial resources. – Neoclassics: Financial liberalization and development of modern big banks and equity market – New structural economics: Optimal financial structure will be different depending on level of development. For low-income countries, small, local financial institutions should be the core of financial structure; and big banks and equity market will play increasingly important role as the firm size and risks increase with the level of development. Fiscal stimulus: – Old structuralism: Keynesian stimulus, using tax and expenditure policies to offset business cycles. – Neoclassics: Ricardian Equivalence, warming against the use of fiscal stimulus. – New Structural economics: Beyond Keynesianism, using public investments to invest in productivity-enhancing, bottleneck-releasing infrastructure projects as countercyclical measures. 16

THE INDUSTRIAL POLICY & GROWTH IDENTIFICATION AND FACILITATION 17

The desirability and failures of Industrial Policy Economic development is a process of continuous process of industrial upgrading and structural transformation. The state should play a facilitating role in the process. Industrial policy is a necessary instrument for the state to play the facilitating role – Contents of coordination will be different, depending on industries. – The government’s resources and capacity are limited. The government needs to use them strategically. The sad fact is that most governments in the developing world used industrial policies but failed, the reason was: – Attempt to develop industries that went against comparative advantage – The firms in the industrial policy’s targeted sectors were non-viable in competitive markets and required government policy supports for their initial investment and continuous operations. – The supports were implemented through price distortions. As a result, planning and administrative allocations were required. – This led to rent-seeking, directly unproductive profit seeking, and soft budget constraints. 18

The existing approaches for industrial development and their drawbacks The existing practices: – Business environment The goal is to introduce a whole set of the first-best institutions The issues are: – The government may not have the capacity to introduce all those changes – The first-best institutions may be different at different stage of development – No identification of industries with latent comparative advantages and no compensation for the first movers – Growth Diagnostics The goal is to remove binding constraints The issues are: – Binding constraints are endogenous to industries – It relies on survey of existing firms. Many of them may be in industries where the country has no comparative advantages. – No firms will be in the new industries that the countries have latent comparative advantage Aim before fire: For an industrial policy to be successful, it should target sectors that conform to the economy’s latent comparative advantage: – Firms will be viable and the sectors will be competitive once the government helps the firms overcome the coordination and externality issues But how to pick the sectors that are the economy’s latent comparative advantages 19

What Can Be Learned From History Historical experiences show that successful countries’ industrial policies, in general, targeted dynamic industries in successful countries with a similar endowment structure and somewhat higher per capita income: – Britain targeted the Netherlands’ industries in the 16 th and 17 th century, its per capita GDP was about 70 % of Netherlands’. – Germany, France, and USA targeted Britain’s industries in the late 19 th century, their per capita income were about 60 to 75 % of Britain’s per capita GDP – In Meiji restoration, Japan targeted Prussia’s industries, its per capita GDP was about 40% of Prussia’s. In the 1960s, Japan targeted USA’s industries, its per capita GDP was about 40% of USA’s per capita GDP – In the 1960s-1980s, Korea, Taiwan, Hong Kong, and Singapore targeted Japan’s industries, their per capita income was about 30% of Japan’s per capita GDP – In the 1970s, Mauritius targeted Hong Kong’s industries, its per capita income was about 50% of Hong Kong’s. – In the 1980s, Ireland targeted information industries, its per capita income was about 45% of the USA’s. – In the 1990s, Costa Rica targeted memory chip assembly and testing, its per capita GDP was about 40% of that of Taiwan, which was the main economy in this sector. Unsuccessful industrial policies in general target industries in countries where their per capita GDPs were less than 20 per cent of those targeted countries. A new approach for industrial policy: Growth identification and facilitation 20

Step 1: Identifying sectors with latent comparative advantage Find dynamic growing countries with a similar endowment structure and with about 100% higher per capita income. Identify tradable industries that have grown well in those countries for the last 20 years as the potential targets of industries for upgrading or diversification Similar to Hausmann’s idea of jumping to nearby trees, but easier to implement Consistent with FDI research suggesting technology transfer is easier when domestic and foreign firms are closer to each other on the technological frontier (Blomstrom, Kokko)

Step 2: Removing constraints for existing firms….How? See if some private domestic firms are already in those industries (of which may be existing or nascent). Identify constraints to quality upgrading, further firm entry, and reduction of transaction costs (hard and soft infrastructure). Take action to remove constraints Methods: – Value-chain analysis – Growth Diagnostics (Hausmann, Rodrik, and Velasco (2008)) – Investment Climate Assessments (World Bank) Successful Examples – Chile: wine – Ecuador: cut flowers

Step 3: Seek FDI or organize New Firm Incubation Programs In industries where no domestic firms are currently present, seek FDI from countries examined in step 1, or organize new firm incubation programs. Famous examples of FDI: – garment sector in Bangladesh – Textile industry in Mauritius – Memory chip assembly and testing in Costa Rica – Electronics and other consumer products in China – Information industries in Ireland – Laura Alfaro and Andrew Charlton, in a paper in the Journal of International Economics, show that: Many countries promote FDI selectively Targeted sectors grow faster Famous example of incubation programs: – Taiwan-China’s Hsingchu Science-based Industrial Park for the development of electronic and IT industries – Fundación Chile’s demonstration of commercial salmon farming

Step 4: Scale up private firm’s self discovery In addition to the industries identified above, the government should also pay attention to spontaneous self discovery by private enterprises and give support to scale up the successful private innovations in new industries Examples – India’s information industry – Ethiopia’s cut flower exports – Peru’s asparagus exports 24

Step 5: Create zones or industrial parks, and encourage industrial clusters In countries with poor infrastructure and bad business environment, special economic zones or industrial parks may be used to overcome these barriers to firm entry and FDI and encourage industrial clusters. Examples: – China’s special economic zones – Mauritius’ export process zone Enormous Increase in Number of Zones World Wide: from 29 in 1975 to 3500 in 2006! – The zones will be successful only if the industries targeted by the zones are consistent with the comparative advantages of their economies

Step 6: Provide limited subsidies to compensate for externalities The need for subsidizing pioneer firms – Information externality of success and failure – Asymmetry of gain of success and loss of failure The governments in developed countries compensate pioneer firms by: – Patents – Supports for basic research – Mandate – Government procurement – Except for patents, the other supports are sector specific The government in a developing country can compensate pioneer firms in the listed identified in step 1 with – Tax incentives for a limited period – Direct credits for investments – Access to foreign exchanges As the government’s support is only to compensate for information externalities, the support can, and should, be limited both in magnitude and time. 26