Parneet Sachdev
The relationship between innovation, research freedom, and economic growth has shaped modern history. The mid-20th century offers a striking example: the Soviet Union’s technological feats, such as the launch of Sputnik in 1957 and sending Yuri Gagarin into orbit in 1961, momentarily convinced many observers that centralized planning might surpass market-driven economies. Influential economists like Paul Samuelson speculated that the USSR would soon outgrow the United States(Samuelson, Economics, 1961; Lange, On the Economic Theory of Socialism, 1936).
Yet, by the late 1980s, the Soviet experiment collapsed, just as the digital revolution was transforming global industry. Despite investing heavily in scientific research—Khrushchev’s creation of Zelenograd, a planned high-tech district near Moscow, being a case in point. Its downfall stemmed not from a shortage of talent, but from rigid institutions that stifled experimentation and local problem-solving. By contrast, Silicon Valley in the United States flourished on decentralized networks of entrepreneurs who iterated rapidly, pivoted easily, and cross-pollinated ideas across firms and sectors (Vedovato& Reinert, “Lessons from Soviet Innovation Failures,” Journal of Economic History, 2010).
CENTRALIZATION VS LOCAL KNOWLEDGE
Friedrich Hayek’s argument, that the problem with central planning lay not in processing information but in acquiring dispersed, tacit knowledge, proved prophetic. Though state investment fuelled heavy industry and basic science, the absence of flexible institutions, open competition, and property rights created systemic stagnation. As the USSR’s GDP per capita plateaued, Western economies surged, powered by innovation-driven sectors like computing, aerospace, and telecommunications.
Data from the World Bank shows that the Soviet Union’s GDP growth rate declined from an average of 5.6% annually between 1950 and 1970 to below 2% in the 1980s, while the US grew at an average rate of 3% over the same period, driven by the information technology sector (World Bank, World Development Indicators, 2022).
Japan’s meteoric rise in manufacturing during the 1970s and 1980s was built on dense networks of coordination and shared knowledge, facilitated by cross-holdings between suppliers and manufacturers. Yet by the early 1990s, Japan’s grip weakened as global innovation shifted toward software-driven models.
Japan’s automotive industry achieved a productivity advantage of nearly 17% over US competitors by 1980, largely through process innovations such as just-in-time production and computer-aided design (OECD, Industrial Productivity Database, 1985). However, by the late 1990s, Japanese firms had failed to transition from hardware optimization to software ecosystems, while US firms like Microsoft and Intel led the computing wave (Mansfield, “Japanese R&D: Process vs Product Innovation,” Research Policy, 1986).
Similarly, Western Europe’s post-war boom depended on mass production and American technologies, which European countries adopted efficiently. By the 1970s, however, Europe’s economic growth slowed as innovation-driven sectors became more crucial. France’s indicative planning system and Italy’s state-owned enterprises, once engines of recovery, became bureaucratic bottlenecks unable to accommodate rapid change. Southern Europe’s prolonged stagnation in the 1980s and 1990s—dubbed the “two lost decades”—highlighted the risks of over-coordination (OECD Economic Surveys, 2002).
INNOVATION AT THE FRONTIER REQUIRES FREEDOM
Technological revolutions follow a discernible pattern. Early stages of industrial development benefit from coordinated catch-up efforts. The steam engine, steel production, and mass assembly lines required standardized processes and state intervention. But as innovation pushes into uncertain territory—whether computing in the 1990s or AI today—the rigidities of centralized control become liabilities.
Entrepreneurs thrive in the long term, not by perfecting known products but having the freedom that fosters exploration of unknown spaces.
Consider the evolution of artificial intelligence. From 2019 to 2024, large language models (LLMs) scaled up by more than 10,000-fold, yet their performance on advanced reasoning benchmarks remains limited. On the AI Reasoning Challenge (ARC), top models score only around 5%, while more flexible approaches like program synthesis or few-shot learning exceed 20% (AI Benchmark Reports, 2024).
Hans Moravec’s observation holds even now: tasks effortless for humans, such as recognizing complex environments, remain difficult for AI. Despite massive datasets and computing power, models lack sensorimotor experience—the embodied understanding humans acquire through trial and error (Moravec, Mind Children, 1988).
A hypothetical LLM trained exclusively on texts from 1633 would have confidently asserted that Earth is the universe’s center. Similarly, a model trained only on 19th-century literature would deny that humans could fly. This underscores the danger of centralized AI approaches that prioritize statistical consensus over exploratory innovation.
CONCENTRATION OF POWER THREATENS INNOVATION
Today’s biggest tech firms are increasingly dominant. Microsoft’s alliance with OpenAI controls approximately 70% of the commercial LLM market, while Nvidia supplies nearly 92% of GPUs critical for AI training (IDC Report, 2024). Alphabet, Amazon, and Meta also invest heavily in AI start-ups, creating ecosystems where incumbent firms shape access to computing infrastructure and data.
Economists Germán Gutiérrez and Thomas Philippon attribute this concentration not to natural scale advantages, but to entrenched lobbying and regulatory capture. Extended patents, sector-specific licenses, and restrictive employment clauses limit competition, dampening labor mobility and hindering the flow of tacit knowledge essential for innovation (Gutiérrez & Philippon, “Investment-less Growth,” Quarterly Journal of Economics, 2022).
China’s story mirrors this trajectory. Recent policies have tightened control: licensing favors state-backed conglomerates, antitrust enforcement is selective, and loyalty tests have replaced merit-based evaluation. Patent counts, rather than breakthrough research, serve as bureaucratic benchmarks. As a result, innovation shifts from frontier research to political signaling (McKinsey Global Institute, China Technology Review, 2023).
While China’s talent pool remains vast, regulatory uncertainty pushes firms to invest more in political networks than in research and development.
LESSONS FOR INDIA: PRESERVING INNOVATION AMID RAPID GROWTH
India stands at a crossroads in its technological development. With a young population, rising digital adoption, and a burgeoning start-up ecosystem, the country has the potential to emerge as a global innovation hub. Yet, lessons from history and contemporary experiences must be learnt.
India’s regulatory landscape must avoid becoming overly prescriptive or bureaucratic. Policies that encourage experimentation, such as relaxed labor mobility rules and simplified licensing for research-oriented start-ups, are crucial. The Startup India initiative and Digital India programs, brainchildren of Prime Minister Modi, have laid early groundwork by easing compliance and providing infrastructure, but deeper reforms—particularly in intellectual property regimes and data-sharing frameworks—are necessary to prevent monopolistic barriers from taking hold (Department for Promotion of Industry and Internal Trade, Government of India, 2023).
The US experience shows that antitrust enforcement can spur innovation by preventing incumbents from stifling competition. Encouraging partnerships between academia, research institutions, and private firms—through grants, open-source platforms, and public-private consortia—will facilitate the diffusion of tacit knowledge that cannot be easily codified in patents (NITI Aayog, India Innovation Index, 2024).
India’s strength lies in its decentralized talent pool—from Tier-2 cities’ tech start-ups to rural entrepreneurs leveraging frugal innovations. Programs such as the Atal Innovation Mission and state-level incubators, once again fostered by PM Modi, should focus not just on infrastructure but on mentoring networks that enable cross-sector experimentation, particularly in agriculture, healthcare, and renewable energy (Atal Innovation Mission, 2023).
As AI emerges as a defining technological frontier, India must ensure open access to datasets while instituting robust privacy and ethical safeguards. Encouraging the development of AI models that learn from diverse linguistic, cultural, and socio-economic contexts will prevent narrow, consensus-driven models from dominating.
The historical record—from the Soviet Union’s collapse to Japan’s stagnation and China’s centralization—makes one lesson abundantly clear: innovation is not merely a function of capital or talent, but of institutions that embrace uncertainty, encourage competition, and reward experimentation. For nations like India, where growth opportunities are immense, designing flexible, transparent, and meritocratic systems is essential to ensuring that innovation.
The frontier of technology will always be uncertain. Where precedent is thin, bold experimentation is the only path forward. History teaches that openness—not control—remains the most powerful engine of economic transformation.
Views expressed are the author’s own).
Parneet Sachdev, IRS is the Chairman of Real Estate Regulatory Authority and a leading author.