← All publications

From the literature

The strangest product of innovation: information

What does a research activity produce? A molecule? A piece of software? A prototype? A new process?

For Kenneth Arrow, we need to go one step further.

What research fundamentally produces is information.

This proposition seems almost obvious. Yet it has considerable economic consequences.

A very particular commodity

A tonne of steel used by one company cannot be used by another at the same time.

Knowledge works differently.

A mathematical formula, a production method or a scientific discovery can be used simultaneously by many players.

The cost of creating it can be immense. The cost of reproducing it can become very low.

That is the problem.

In his article, Arrow explains that if transmitting information cost nothing, the economically optimal allocation would be to disseminate it as widely as possible.

Once knowledge has been produced, preventing its use reduces the benefits it could generate.

But if everyone can use it immediately and free of charge, the party that paid to produce it can no longer recover its investment.

The information seller's paradox

Arrow also highlights a particularly elegant problem.

To know the exact value of a piece of information, the buyer needs to know that information.

But once it has been revealed to them so that they can assess it, they already have it.

Try selling someone a piece of secret information.

The buyer asks you: "What does it contain?"

You may reply: "Pay and I will tell you."

But how can they determine what it is worth before knowing what it contains?

This problem sets information profoundly apart from ordinary goods.

Value spills over beyond the company

This property is essential when it comes to R&D.

A company may finance a series of experiments and discover that the solution it had in mind does not work.

In accounting terms, the result may look negative.

In economic terms, however, information has been created.

We now know that a technological path is probably a dead end.

This knowledge can avoid further expenditure. It can redirect a research programme. It can make it possible to choose another technology.

Experimental failure is therefore not necessarily an absence of output. It can be a production of information.

A major difficulty for financing

The problem is that the value of this information cannot always be monetised immediately.

A company pays for the experiments today. The benefits may appear much later.

They may also appear in another project. Sometimes they will even benefit other players.

The private investment decision therefore does not necessarily take into account all the value created.

We come back to Arrow's conclusion: private investment in the production of knowledge is likely to fall short of what would be socially desirable.

What this changes for an innovative project

A company preparing an application for public funding therefore has an interest in not presenting only its final product.

It must also describe the knowledge production process.

What do we know today? What do we not know? Which hypotheses are we going to test? Which decisions can be taken thanks to the results obtained? What new knowledge will the programme create?

This way of reading profoundly transforms the way a project is told.

R&D is no longer a mysterious black box between an idea and a product.

It becomes an organised succession of uncertainties and new information. And it is precisely this production of information that the funding is meant to pay for.

Main reference: Kenneth J. Arrow (1962), Economic Welfare and the Allocation of Resources for Invention, in The Rate and Direction of Inventive Activity: Economic and Social Factors, Princeton University Press / NBER, pp. 609-626. The economic reasoning is attributed to Kenneth Arrow. The parallels drawn with contemporary public funding and CHOICE's work are an editorial perspective.