What if the key to understanding sustained economic growth lies not in the accumulation of more machines and factories, but in what we learn while building them? This is the revolutionary insight behind learning-by-doing models-a family of economic theories that transformed how we think about productivity and progress. From Kenneth Arrow’s pioneering work in the early 1960s to Paul Romer’s Nobel Prize-winning contributions, these models reveal how experience, observation, and investment create knowledge that fuels enduring economic expansion.
Table of Contents
The Arrow model: when experience becomes productivity
In 1962, economist Kenneth Arrow introduced a groundbreaking concept that challenged conventional thinking about economic growth. Rather than treating technological progress as something that simply happens over time-like a ticking clock-Arrow argued that workers and firms acquire valuable skills and knowledge through the very act of producing goods and services.
Arrow’s central insight was remarkably straightforward yet profound: productivity improves through hands-on experience, not just through formal research and development. He drew inspiration from empirical observations in the aerospace industry, where engineers noticed that the labor hours required to build each successive aircraft decreased in a predictable pattern. The more planes they built, the more efficient they became-not because of new equipment, but because workers learned better techniques through repetition and problem-solving.
In Arrow’s framework, this learning process is treated as a by-product of investment and production. As firms accumulate capital-building factories, installing equipment, producing output-they simultaneously accumulate knowledge about how to do things better. This knowledge becomes embedded in the production process itself, raising productivity without requiring deliberate innovation efforts.
However, Arrow’s original model faced a limitation. If all learning comes automatically from experience and the labor force remains constant, growth eventually slows down because society invests and produces less over time. Arrow’s vision preceded modern endogenous growth theory by two decades, setting the stage for later economists to build upon his foundation.
The Levhari-Sheshinski model: knowledge as a public good
Building on Arrow’s work, economists David Levhari and Eytan Sheshinski extended the learning-by-doing framework in important ways. Their key contribution centered on recognizing that knowledge created by one firm doesn’t stay locked within that firm’s walls-it spills over to benefit the entire economy.
In the Levhari-Sheshinski model, when a firm invests in training its workforce or developing better production methods, it creates knowledge that acts as a public good. Other firms can observe, imitate, and adapt these innovations at virtually zero cost. Think of how a breakthrough manufacturing technique pioneered by one automaker eventually gets adopted throughout the industry, or how management practices developed at one company spread through business schools and consulting firms to organizations everywhere.
This spillover effect creates a powerful multiplier for economic growth. While each individual firm operates under constant returns to scale, the economy as a whole experiences increasing returns because knowledge accumulates and spreads. One firm’s investment in learning raises not just its own productivity but also the productivity of countless other firms that can tap into this expanding knowledge base.
The model helps explain why economies with vibrant business ecosystems-where firms can easily observe and learn from each other-tend to grow faster than isolated economies. The free flow of ideas and techniques creates positive externalities that amplify the impact of any single firm’s innovative efforts.
The King-Robson model: the power of watching
Ian King and Arthur Robson introduced another fascinating dimension to learning-by-doing theory: learning by watching. Their model emphasizes that firms don’t just learn from their own experiences-they learn by observing the innovations and problem-solving approaches of others.
When one firm successfully implements an innovation-whether it’s a new production technique, an organizational structure, or a business model-it creates a demonstration effect. Other firms witness this success and are encouraged to adopt and adapt similar innovations to their own circumstances. This process works like a contagion, spreading productive knowledge throughout the economy.
Imagine a textile manufacturer that develops an innovative quality control system. As word spreads about the improved defect rates and customer satisfaction this firm achieves, competitors take notice. They study the approach, modify it for their specific needs, and implement their own versions. Before long, industry-wide quality standards have risen, all stemming from one firm’s initial innovation.
The King-Robson study reveals that innovation in one sector has a contagion or demonstration effect on the productivity of other sectors, leading to economic growth across the board. Importantly, their work shows that economies with similar initial resources can end up on different growth trajectories depending on how effectively they facilitate this “learning by watching” process.
This insight has powerful policy implications: governments should foster environments where firms can easily observe and learn from each other-through industry associations, technology parks, business networks, and knowledge-sharing platforms. Investment policies that encourage experimentation and visible innovation can trigger broader growth cascades.
The Romer model: investment as knowledge creation
Paul Romer, whose contributions to growth theory earned him the Nobel Prize in Economics in 2018, developed what became known as “learning by investment.” In Romer’s framework, knowledge creation is a deliberate by-product of investment decisions, not merely an accidental side effect.
Romer’s production function incorporates a public stock of knowledge that depends on aggregate research and development efforts, alongside firm-specific R&D investments. The key innovation is recognizing that knowledge has unique economic properties: it’s non-rival (many people can use the same idea simultaneously without depleting it) yet partially excludable (through patents and intellectual property protections).
Romer stressed that the non-rivalry of ideas gives rise to increasing returns to scale. Unlike physical capital that gets divided among workers, knowledge can benefit everyone in the economy simultaneously. When one researcher discovers a breakthrough, that discovery becomes part of humanity’s collective knowledge base, potentially raising productivity for countless individuals and firms.
Romer’s model features three critical elements that distinguish it from earlier approaches. First, externalities: research efforts by one firm create spillovers that help other firms develop new knowledge. Second, increasing returns in output production: as the stock of knowledge grows, the economy becomes more productive at an accelerating rate. Third, diminishing returns in knowledge production: each additional unit of research investment yields progressively smaller knowledge gains, preventing unlimited growth rates while still enabling sustained expansion.
The model explains why advanced economies with large stocks of accumulated knowledge and robust R&D sectors tend to grow faster than economies starting from scratch. It also clarifies why policies supporting research, education, and intellectual property protection can influence long-term growth rates-something earlier growth models struggled to explain.
From theory to practice: what learning-by-doing models teach us
These four models share a common thread: knowledge accumulation, whether through doing, watching, or deliberate investment, drives sustained economic progress. Unlike older theories that treated technological change as an external factor-something that just happens-these models make growth endogenous, arising from within the economic system itself through purposeful actions and natural learning processes.
For policymakers, the implications are significant. Countries with greater stocks of human capital and more investment in research and development will enjoy faster economic growth rates. This helps explain why some nations advance rapidly while others stagnate despite similar resource endowments.
Investment in education becomes crucial not just for individual worker productivity but for its spillover effects across the economy. When firms invest in training, the knowledge created often spreads beyond organizational boundaries, raising productivity economy-wide. Similarly, policies that facilitate knowledge sharing-whether through industry clusters, research consortiums, or open innovation platforms-can amplify the growth effects of any single firm’s learning.
These models also highlight the importance of an ecosystem where experimentation is encouraged and innovation is visible. Economies that penalize failure too harshly or that prevent firms from observing and learning from each other’s experiences miss out on powerful growth multipliers inherent in learning-by-doing processes.
Perhaps most importantly, these theories suggest that growth is not predetermined by geography or initial conditions alone. Through deliberate policies supporting knowledge creation, human capital development, and knowledge diffusion, societies can influence their long-term growth trajectories in meaningful ways.
What do you think? How might businesses in your industry better harness learning-by-doing effects to boost productivity? What policies would best facilitate knowledge spillovers between firms in developing economies?
Leave a Reply