Day 3

Lecture 2

Growing Entrepreneurial and Innovation Clusters

A lecture by Josh Lerner

Jacob H. Schiff Professor of Entrepreneurship, Harvard Business School; Co-Director, HBS Private Capital Project.

Why venture capital matters, and why governments chase it

Venture activity is both far larger and far more global than it used to be. In 2001 some 80% of investment was in the United States; today the U.S. share is closer to 30–40%, with China and India accounting for much of the growth, and reaching Western levels of venture activity at far lower GDP per capita. Governments have responded with enormous public expenditures: over $5 billion in subsidies for the TSMC facility in Arizona, roughly $3 billion for the Micron semiconductor plant in Gujarat (about 18 months of budget for India's premier technical institutes, on Raghuram Rajan's calibration), and over $50 billion in Saudi Arabia's AI initiative alone.

Why the enthusiasm? Because the innovative footprint of venture is out of all proportion to its size. In the United States only about one in a thousand new businesses receives venture funding, yet half of the companies that have gone public in the last quarter-century were venture-backed, and that subset accounts for roughly 90% of the R&D spending of recent public companies. The pattern holds abroad: venture-backed firms represent about 40–50% of such R&D across emerging markets, and in developing countries outside China they account for roughly 31% of patenting (around 40% in knowledge-intensive industries). Meanwhile corporate research has been a mixed bag, with firms retreating from basic science; estimates suggest venture is three to four times more effective than corporate R&D per dollar.

Why so many programs fail, and what works instead

Everywhere with a successful industry shows the strong hand of government early on, the U.S., China, Israel, Singapore. But many governments have spent lavishly with little return, Europe being the clearest example. Two failure modes recur. First, officials plunge in without understanding the dynamics and chase the trend of the moment: Queensland, Australia poured mining royalties into biotechnology labs, only to watch the good ideas leave for Sydney or San Diego, while the firms actually winning export awards were drone-and-software businesses with a genuine local advantage. Second, and less common, programs get captured. Japan built the world's fourth-largest venture industry on subsidies in the 1990s, then watched it vanish once the subsidies stopped, because tax and employment policy still made entrepreneurship hard.

The constructive agenda rests on three principles. Set the table: get the institutional environment right first, exit markets (second-tier exchanges with real rule of law), flexible labor markets, working technology transfer, sensible taxes, and reduced penalties for failure (Italy's old bankruptcy register discouraged risk-taking; Singapore's "Phoenix Award" celebrated meritorious failure). This work is unglamorous and resisted by vested interests, but essential. Let the market lead: use matching-fund requirements so participants stake their own money, and welcome foreign investors where local experience is thin. Iterate and be patient: clusters take decades, the modern U.S. industry began in 1946 but did not take off until the late 1970s.

Evidence at a glance

• Young firms drive net job growth. Across developed and developing economies, firms five years old or less account for the bulk of net new jobs.
• Venture is high-powered money. It is estimated three to four times more effective than corporate R&D per dollar in generating innovation.
• A tiny slice, an outsized footprint. About 0.1% of U.S. businesses get venture funding, yet venture-backed firms are roughly 90% of the R&D done by recent public companies; at the frontier, Sequoia estimates about 65% of its companies end with only nominal returns.
• Europe's gap is structural. U.S. venture activity is about four to one versus Europe; the U.S. has around nine companies worth $1 trillion or more founded in the last 50 years, Europe none worth more than 100 billion euros.
• Skills are about opportunity-spotting. Two of a few hundred Harvard College entrepreneurship students became billionaires in their early 20s, success tracked adaptability and vision more than formal credentials (cf. Jack Ma).

Implications for middle-income countries

• Target where you have an edge. Not everyone needs a foundational AI model; the application layer, biotech, or clean tech may fit better, and let real, if early, local promise guide the choice.
• Small nations must think regional and international. Domestic markets alone are too thin; the UAE's success came from attracting firms with "a foot in different camps," not from funding only nationals for a national market.
• Draw on the diaspora. Singapore and Israel leaned on expatriates abroad for capital and expertise rather than relying solely on home-grown investors.
• Exploit the follower's advantage. Emerging-market entrepreneurs succeed by adapting what worked elsewhere to local conditions, Careem localizing ride-share for the MENA region is one example.
• Learn from documented cases. Two decades of experience, successes and failures alike, now codified in OECD volumes and country studies, give policymakers far more to learn from.

Lecture 3

Spatial Creative Destruction and Regional Development

A lecture by Leonard Wantchekon

James Madison Professor of Political Economy and Professor of Politics and International Affairs, Princeton University.

Science, productivity, and the roadblock of state capture

Following Abdus Salam's line that "development is science with some noise," the lecture argued that science is the ultimate driver of progress, the reason life expectancy has climbed from roughly 35 two centuries ago to the low-to-mid 70s today. Science works through two mechanisms: productivity (a country's technological competence, in Nelson's phrase, is the basic factor controlling it) and, often overlooked, state capacity (technology lets states deliver services, a passport in Benin can now be issued in a day, and lets citizens act for themselves). And, as the innovation-and-competition literature shows, creative destruction is more likely in a democratic, competitive environment, with democracy most growth-enhancing near the frontier.

The obstacle is political. As work by Aghion and co-authors, and by Acemoglu and Robinson, has shown, political elites can block progress when dynamic innovation threatens their interests. Data from Italy make the mechanism concrete: larger firms are more likely to be politically connected, and connected firms are older and less innovative, so incumbents entrench themselves through connection to deter entry. But much of the developing world faces something worse, state capture, where firms and militaries own the economy and suppress competition through policy and repression alike (the Gupta brothers' capture of a South African ministry is one example). A study of local elections in Benin distinguishes indirect capture (a promise of policy) from direct capture (appointing the bureaucrat who will implement it). Direct capture is the more prominent form, and, counterintuitively, it rises when elections are more competitive: fearing promises won't survive a change of government, firms buy the official who stays regardless of who wins.

Decentralized safe havens

If central power is captured, where can creative destruction begin? The answer is decentralized safe havens: pick one place or sector where the rents elites seek are low (so they ignore it) and where people are eager for innovation, prove the model there, and let it spread through social learning and local externalities of proximity, culture, and complementarity. The parallel is China's special economic zones, but with science, not extractive industry, at the core, development as externalities moving between regions, not a top-down plan.

The idea grew out of a personal shock. Returning to his mother's fertile hometown, the speaker found a once-thriving agricultural market dead, because a single bridge to market had collapsed, after which young people left for moto-taxi work in the city. The data confirmed the pattern: across Sub-Saharan Africa, a 10% increase in soil suitability is associated with a 2% increase in poverty, because the best roads sit on the worst land. Colonial roads reached mines, not fertile soil, and since independence countries have merely repaved those same north–south routes, leaving weak market integration.

Fertile river regions are natural safe havens (some support four cropping cycles a year) but they need a research station to unlock that potential. Research infrastructure endures: a French oil-palm research center founded in 1922 still shapes its region's development. Frontier science is feasible on the continent (Ethiopia's space program for remote sensing, and teff, a drought-resistant grain now recognized globally as a superfood), and where it takes hold it becomes hard to reverse. Brazil's EMBRAPA (1973) is the model: research placed close to local ecology, a talent pipeline through a U.S. university, and, by recent estimates, a contribution of between 42% and 78% of total factor productivity growth.

Evidence at a glance

• Fertile land, more poverty. Across Sub-Saharan Africa, a 10% rise in soil suitability is associated with a 2% rise in poverty, because market access, not soil, drives outcomes.
• Competition can worsen capture. In Benin, direct state capture (appointing implementing bureaucrats) is the dominant form and increases as elections become more competitive.
• Research infrastructure pays off. EMBRAPA (1973) accounts for an estimated 42%–78% of Brazil's total factor productivity growth; a 1922 oil-palm center still marks its region a century on.

From despair to a plan

• Build the missing anchors, with extension. The gap is rarely roads alone but the agronomic research station that raises productivity, and, as with EMBRAPA, the technical assistance and cooperatives that must accompany the science.
• Connect farms to markets. Link high-productivity rural areas to cities and to Nigeria's roughly 280-million consumer market, and add logistics to cut waste.
• Put AI to work. It can forecast prices and cut waste in agriculture, speed learning from research stations to farmers, and (via a Wikipedia-style platform) recover African history and languages otherwise being lost.
• Institution-build directly. The African School of Economics (opened 2014) is adding engineering, a research hub, and a private "agricultural complex" of experimental farms and cultural tourism, a roughly $44 million venture.