RDP 9213: The Impact of Financial Intermediaries on Resource Allocation and Economic Growth 2. Growth Through Knowledge Accumulation

The traditional neoclassical view of economic growth is associated with the work of Solow (1956 and 1957). He argued that differences in capital-labour ratios were the key determinant of cross-country differences in the level of per-capita incomes. In Solow's model, the rate at which per-capita income increased in the steady state was explained by exogenous technological progress. This exogeneity assumption meant that there was no consideration of how or why technological progress took place. The new generation of growth models explicitly address these issues by modelling the technology and knowledge accumulation processes. In so doing, they add considerably to our understanding of the economic growth process.

In these models, knowledge is defined in various ways. In the Romer (1989) model, knowledge is equated with the number of blue-prints for intermediate goods that have been developed. In the model presented by Lucas (1988), knowledge is defined as the stock of human capital. In other models it is defined as production knowledge gained from learning-by-doing and in others as the level of technological sophistication. In the discussion that follows, the term knowledge is used in a generic sense to include each of these forms of knowledge.

Romer assumes that the economy has three sectors: a sector producing final goods, a sector producing differentiated intermediate goods and a research sector producing “blueprints”. The final good is produced using unskilled labour (L), skilled labour (H₁) and all possible varieties of intermediate goods that are currently available. The production function for final goods, which exhibits constant returns to scale, is given by:

where x(i) is the input of variety i of the intermediate good. An important characteristic of this production function is that output increases when the number of varieties of intermediate goods increases, even if the total input of intermediate goods is held constant; that is, there are returns to variety in intermediate goods.

The intermediate goods are produced with constant marginal cost with capital being the only input. The capital stock increases through time as consumption is foregone. Before an intermediate good can be produced, it must be developed by the research sector. Once developed, the firm that wishes to produce the intermediate good buys the design from the research sector. The increase in the number of new designs (which is equated with an increase in knowledge) depends upon three factors: the stock of existing knowledge (A), the amount of skilled labour used in the research sector (H₂) and the productivity of that labour in creating new varieties (δ). The increase in the number of new varieties is given by:

This production function exhibits increasing returns to scale. However, producers of the blueprints are required to pay only for the skilled labour; they are able to use the existing stock of knowledge (A) for free. This allows the existence of a competitive research sector.

Increases in the stock of knowledge have two important effects on the economy. First, the increasing stock of knowledge makes it easier to produce new knowledge (that is, new varieties) and second, it increases the number of varieties of intermediate goods available to final good producers. This in turn increases output.

Consumers are assumed to have a standard intertemporal constant elasticity utility function given by:

Each consumer is assumed to be endowed with fixed quantities of unskilled and skilled labour. In each of the industries profits are zero. Finally, equilibrium is assumed to exist in the market for skilled labour; that is H₁+H₂ = H, where H is the stock of skilled labour in the economy.

Romer shows that the equilibrium growth rate in this economy (g) is given by:

The growth rate is increasing in the stock of human capital (H) and in the productivity of human capital in creating new knowledge (δ). The more efficient is the society in using its existing knowledge to create new knowledge, and the more resources it has to devote to knowledge creation, the faster will be its growth rate. The growth rate is negatively related to the rate at which individuals discount the future (ρ). A higher discount rate leads to reduced capital accumulation and thus a reduction in the production of intermediate goods. The growth rate is increasing in the elasticity of intertemporal substitution (1/σ); that is, as consumers become more willing to substitute consumption between periods, the growth rate increases as consumers are prepared to forgo consumption today for higher consumption tomorrow.

Romer's model has many interesting implications for economic growth and the role of government in encouraging growth. Unlike the Solow growth model, the preference parameters (ρ and σ) affect the steady-state growth rate. More patient societies will grow faster. Romer also shows that since researchers are not able to capture the full returns to society of their innovations, the growth rate of the economy will be less than socially optimal. Government subsidisation of research and development thus may be desirable. Further, the model suggests that policies aimed at increasing the efficiency with which the economy translates its existing resources into faster knowledge accumulation are likely to speed the steady-state growth rate.

The work of Romer and others has been important in re-emphasising the role of knowledge accumulation in the growth process. It emphasises that the decision to create knowledge is one made in response to market forces. It also emphasises that there are important positive externalities for society from the creation of knowledge, and thus left solely to the market, there may be insufficient accumulation of knowledge. The Romer model, in ignoring financial markets, is typical of models in this framework. Savings, in the form of foregone consumption, are assumed to be costlessly and efficiently transformed into capital to be used in the production of intermediate goods. Financial markets are implicitly assumed to be working in the background allowing individuals to substitute consumption between periods. In addition, knowledge creation happens the moment the resources are devoted to research and there is no risk involved in the innovation process. Since knowledge creation occurs instantaneously it does not have to be financed through financial markets.

Clearly, the attention given by governments to the operation of the financial system suggests that financial markets should not be pushed into the background when considering economic growth. Given that research and development and knowledge accumulation are risky activities and that the “inventor” may have a considerable information advantage over the lender, questions concerning the incentives of financial intermediaries become important. In terms of the above model, a more efficient financial system, through say its ability to improve the screening of lending projects, may increase the productivity of existing knowledge and human capital. This can be represented by an increase in the parameter δ. As equation (4) shows, increases in δ lead to faster economic growth.

More generally, the financial system, through its roles of screening proposed lending projects, facilitating intertemporal substitution in consumption and investment and influencing the degree of uncertainty in the economy, has an important influence on the rate of economic growth. These issues are considered in more detail in the following section.