Nelson (1959) - The Simple Economics Of Basic Scientific Research

Jump to navigation Jump to search
Has bibtex key
Has article title The Simple Economics Of Basic Scientific Research
Has author Nelson
Has year 1959
In journal
In volume
In number
Has pages
Has publisher
©, 2016


  • Nelson, R.R. (1959), "The Simple Economics of Basic Scientific Research", Journal of Political Economy 67, 297-306. pdf


The article does not have an abstract

Summary of the Paper

Questions of the Paper

The paper addresses the questions:

  1. How much should be spent on basic scientific research?
  2. Why does the private sector not spend the socially desireable amount?

Social value, cost and profit

  • The social value (of basic research) is defined as: The properly discounted flow of benefits arising from a given expenditure on scientific research
  • The social cost (of basic research) is defined as: The discounted flow of benefits that we are deprived of by allocating resources to scientific research rather than other activities.
  • The social profit is defined as the difference between the social value and the social cost. A society should allocate resources to basic science to maximize social profit.

When the marginal value to society exceeds the marginal value to individual who paid for it, the allocation of resources will not be optimal. Thus, with a positive externality to basic research, private markets will fail and there is a rationale for collective support of production of research, such as through positive public policy.

Science and its Benefits

"It is sometimes argued that most of our great social
 and political problems would simply evaporate if all 
 citizens had a scientific point of view"
                                    -R. R. Nelson 1959

The benefits to science are not just the inventions generated, but that it makes for better citizens! Many scientists believe that science is an ends unto itself (the pursuit of knowledge), however, an economist would consider the benefit as the increase resulting from scientific research in the value of the output flow that society can produce.

Scienfic research is the human (sic) activity directed towards the advancement of knowledge. Knowledge is either:

  • Facts or data observed in reproduceable experiments
  • Theories or relationships between facts (usually equations).

Invention is defined as the human activity directed toward the creation of new and improved practical products and processes. Many inventions occur through systematic effort to achieve a stated goal, but some result from research directed toward a different goal (or no goal other than the advancement of knowledge). In the activity of invention there are many paths to choose from - the greater a researcher's relevant knowledge the more likely he will eventually find a satisfactory path. Therefore the greater the underlying knowledge, the lower the expected cost of a particular invention. Inventive activity will only be undertaken if the (private) expected revenue exceeds the (private) expected cost. If the benefit is high, high cost project can be sustained (ignoring risk aversion). To the extent that invention is predictable and to the extent that a private firm can capture (through the market) the full value of the invention to society, applied research for private profit will be conducted at the socially optimal level.

However, not all scientific research is applied: there is a continuous spectrum ranging from basic research to applied research. Note that having loose goals for basic research is rational, given its great uncertainties, and this strategy permits a higher payoff per research dollar than if goals were closely defined. The goal of any applied research is closely constrained, whereas the goal of basic research is adaptive.

Pasteur's Chickens

The paper relates the story of Pasteur's chickens on p301. Pasteur was experimenting with chicken cholera, and accidentally innoculated a batch with a weak strain. They got ill and then recovered. Pasteur did not want to waste chickens and so later gave them a fresh culture - which should have killed them - but they remained healthy. Pasteur then shifted his research to working out why.

Basic Research and Private Profit

Basic research is likely to generate substantial positive externalities. Private profit motives will therefore be insufficient to achieve social optima. Significant advances in scientific knowledge are often not generally applicable and do not result in patents. Often the knowledge is most valuable as a key input to other research - for this reason natural laws and facts are not patentable.

The degree to which a firm can capture the benefits of more basic science depends on how broad or narrow its technological base is. Narrow base firms should only conduct applied research directly in their area, whereas broad firms are in a better position to capture the more 'random' outcomes of more basic research. For this reason, firms that support basic research "have their fingers in many pies". It is not just the size of the firm that matters, it is the breadth of its base, which usually entails a diversified set of products. A wide base does not imply monopoly in a particular market (or vice versa).

However, firms in competitive environments will be less willing to engage in research which does not quickly result in a patent, as to extract value from the research they would otherwise have to keep it secret and risk losing the advantage of their findings to their competitors. Monopolists are perhaps better able to 'store' knowledge until it is commercially applicable and so better positioned to invest in research that might need to be stored. Note that this is in stark contrast to Arrow's arguments. One solution to this 'public goods' like problem is for industries to create cooperative research laboratories.

There are two other factors that might prevent research from being undertaken in competitive markets:

  1. The long lag between initiation of a project and commercial value realization might be prohibitive for firms that value short term survival and profits over longer term objectives. Possibly monopolists have lower discount factors, and almost certainly governments should (again c.f. Arrow)
  2. Firms (and individuals) are risk averse and projects are risky. Firms without the economic resources to spread risk (by diversification across projects) will value projects less than its social value, even without externalities.

Is Current Social Policy Adequate

The argument is that it doesn't exceed the social optimum is as follows.


  1. Basic research is a homogenous commodity - the public is indifferent as to whether it is done publicly or privately
  2. The marginal cost of public research is no greater than that of private research
  3. Private laboratories operate where marginal revenue equals marginal cost

Then the finding that private laboratories do basic research at all is evidence that we should increase our expenditure.

Save for tax laws, public policy has not shifted the marginal cost curve, but has resulted in shifts along the curve. Externalities still exist at the margin.

However, basic research is not a homogenous commodity, and so the concept of relative marginal cost (between public and private) is not clear, so an airtight statement based on welfare economics is not possible.

Policy Implications

There is a basic contradiction between the conditions necessary for efficient basic research (no constraints and full dissemination) and full appropriation of the gains from sponsoring basic research in a competitive economy.

"If we advocate that basic research be increasingly undertaken by business (which should be profit motivated), we must accept the growth of large firms with a wide technology base, with virtual monopolies in several markets." Otherwise concious social policy is the way to go!

Recogition of the following points is important:

  1. There is a classic externality problem. The research undertaken is often of little value to the firm that bears the cost but of great value elsewhere (allocative efficiency), and research often cannot be quickly patented (discounting and capture problems). Solutions include industry coops, contract research firms, and universities.
  2. Incentives in profit maximizing firms lead to secrecy and so static inefficiency. The marginal cost of using knowledge that already exists is zero, and knowledge is non-rivalous, so knowledge should be contributed to a common pool. But if this is implemented, the incentives to create knowledge will be reduced, leading to dynamic inefficiency.
  3. Universities need funding (a good pitch!), to prevent scientists from being drawn away from basic research into applied research. A university's comparative advantage lies in basic research.