Knowledge Externality in Production

A pharmaceutical company spends billions developing a new drug. Once it launches, competitors observe the strategy and adapt their own research. An AI lab publishes a paper on neural networks; rival labs cite it, learn from it, and build on it without paying a licence fee. This is knowledge externality: one firm’s investment in ideas raises the productivity of others without compensation. Because the discoverer cannot capture all the benefits, she underinvests. This is a classic market failure, and it justifies public support for R&D.

The non-rival nature of ideas

To understand knowledge externality, start with what makes it different from other goods.

A car is rival: if I own it, you do not. If I drive it, you cannot. So I can sell it or hire it out, and the buyer pays because they value the use. Scarcity enforces property rights.

An idea is partly non-rival. If I know how to synthesise a chemical compound, your knowing it does not prevent me from knowing it. Both of us can use the idea simultaneously. Moreover, ideas are non-excludable: once I publish a discovery or a rival observes my method, I cannot easily prevent them from using it. I cannot post a guard around the knowledge.

This creates a fundamental tension. A firm considering whether to invest £100 million in R&D asks: will I recoup this cost? If the moment I invent something, competitors learn the method and copy it without paying, my returns collapse. I receive no compensation for the benefit I created for them. If I cannot extract payment, I under-invest.

The private-versus-social return gap

Economists measure this with two numbers. The private return on R&D is what the inventing firm captures: revenue from selling the product, less R&D costs and imitation losses. The social return includes the private return plus the benefit flowing to others: rivals who use the idea to improve their own products, consumers who benefit from later innovations that build on this one, and the general productivity boost the economy enjoys.

Estimates vary, but most research finds the social return is two to four times the private return. A firm that invests in R&D expecting a 10% return might be creating a 30% return to society. The difference is the externality: value created but uncompensated.

Because the private return is too low, the firm invests less than is socially optimal. Some projects that would benefit the world (a cancer vaccine, an efficient solar cell) are not pursued because the firm cannot capture enough of the value. From society’s perspective, this is a loss.

Mechanisms of spillover

How do these spillovers actually occur?

Reverse engineering and imitation. A competitor observes the innovator’s product, learns the recipe, and manufactures a knock-off. If patents are weak or expire quickly, imitation is fast and nearly free.

Labour mobility. An engineer works for the innovative firm, learns its methods, and moves to a rival firm, taking knowledge with her. Non-competes can slow this, but they cannot stop it entirely.

Publication and disclosure. Academic labs and some firms publish results. Others read the papers and build on them. Even patented inventions must be disclosed in detail, so competitors can study the patent and engineer around it.

Supplier and customer relationships. A manufacturer improves its processes; customers see the better quality and demand similar improvements from their other suppliers. Suppliers copy the innovation.

Industry standards and shared infrastructure. When the semiconductor industry develops a new manufacturing technique, multiple chip makers adopt it, spreading the benefit.

None of these is a market transaction. The original innovator does not charge the imitator, the engineer’s former employer does not invoice the new one, and the researcher does not bill readers of her paper.

Why this matters for growth

Knowledge externality sits at the heart of modern growth theory. If firms under-invest in innovation because they cannot capture the full social return, then the economy’s long-run growth rate is sub-optimal. Public investment in R&D—universities, government labs, infrastructure for basic science—can partly correct this.

Countries that subsidise or directly fund R&D (the US, Switzerland, Germany, South Korea) tend to have higher productivity growth than those that rely entirely on private firms to innovate. This is not coincidence; it is the logic of externality.

Moreover, if one nation invests heavily in R&D and others free-ride on its discoveries, the free-riders gain without bearing the cost. This creates a dynamic where each country under-invests from a global perspective, even if one is wiser than another.

The patent solution—and its limits

Governments have long tried to internalise knowledge externality through patents. Grant a firm exclusive rights to use an invention for 20 years, and the firm can charge monopoly prices, capturing much of the social value. This incentivises investment: the firm now expects a high private return and will invest accordingly.

Patents work, but imperfectly. They are slow to obtain (years of legal work), expensive, and hard to enforce across borders. They protect the firm’s final product but not the underlying ideas; rivals can often invent around a patent by using a different method. And patents create a trade-off: the inventor earns high profits, but monopoly prices reduce access, and the threat of litigation can chill downstream innovation.

Most countries find that patents alone are insufficient. They combine patents with other policies:

• Tax credits for R&D: reduce the cost of innovation to the firm, partly internalising the benefit.
• Direct government funding: universities and national labs conduct basic research, whose spillovers are immediate and broad.
• Intellectual property protection: enforce patents, copyrights, and trade secrets so firms can retain returns.
• Education and skills: invest in human capital so the economy can absorb and build on innovations.

Heterogeneity across sectors

Knowledge spillovers are not uniform. In semiconductors and software, imitation is rapid and cheap; spillovers are fast and large. In pharmaceuticals, patents are stronger and reverse engineering is slower; spillovers are smaller but the firm can recover some costs through pricing before generics launch. In agriculture, knowledge spreads through seed saving and farmer networks; markets barely exist for ideas, so spillovers are near-total.

This heterogeneity implies policy should vary. High-spillover sectors (basic science, agriculture) warrant more direct public funding. Low-spillover sectors (pharmaceuticals, software before open-source) may thrive on patents alone.

See also