Smaller World, Bigger Issues

On the other hand, in his seminal book in 1942, Schumpeter challenged this paradigm and defined competition as a disequilibrium process in which firms face a constant pressure of differentiating themselves from the others in order to survive (Mazzucato, 2000). The market power that firms acquire as a result of being different from the other firms, or the hope of acquiring it, causes them to engage in innovative activities or to adopt new technologies, and hence, constitute the engine of technological progress. Therefore, there is a positive relationship between market power and innovative activities (Van Cayseele, 1998). Schumpeter asserts that, apart from being unrealistic, perfect competition is also a suboptimal phenomenon because highly concentrated market structures are more favorable to technological progress and economic growth. The long-term welfare growth achieved as a result of high innovation rate in imperfectly competitive markets more than compensates the welfare losses resulting from the short-term allocation inefficiencies (Kamien and Schwartz, 1975).

In essence, studies on the relationship between market structure, innovative activities, and economic growth are the reflections of the debate between these two paradigms on the dynamic efficiency of market structures. This chapter reviews the literature and presents the ideas on the issue in a historical flow, without any intention of being exhaustive. The second section presents the models that try to explain economic growth with perfect competition, and the third section with imperfect competition. Finally, the fourth section concludes.

2 Early Models of Economic Growth with Perfect Competition

Arguably, the model developed by Solow and Swan separately in 1956 has been the pioneering and the most influential work in the economic growth literature. Apart from drawing the route of the literature, it has also been used as a benchmark to evaluate the models that followed. The Solow-Swan model focuses mainly on how the growth rate of output per capita is related to the accumulation of production inputs, such as capital, labor, and “knowledge” (technology).

In the model, the total amount of output that can be produced in an economy using the available levels of inputs is represented with an aggregate production function. Each period, economic agents save an exogenously given, fixed fraction of this output to make capital investments and consume the rest. The amount of capital investment and depreciation determines the next period level of capital. Thus, the dynamics of capital accumulation is endogenized in the model. However, the growth rates of other factors, namely, labor and knowledge, are exogenously given.

The Solow-Swan model predicts that the economy will eventually converge to a “balanced-growth path” on which all the variables grow at a constant rate. The growth rate of output and capital are determined by the growth rates of labor, and technology, which are exogenous to the model. However, the landscape completely changes when the analysis is conducted with per capita variables. In this case, the model converges to a steady state in which output per capita and capital–labor ratio remain constant regardless of the saving rate, namely, per capita growth eventually ceases. This result means that capital accumulation cannot be the engine of long-term sustainable per capita output growth.

With the growth accounting approach Solow developed in his 1957 study, he decomposed growth according to its sources. In this approach, the part of the growth that cannot be explained by the accumulation of production factors is attributed to the increase in total factor productivity and referred to as the “Solow residual”. This residual is regarded as representing the contribution of technological progress on economic growth. The US Bureau of Labor Statistics used the growth accounting approach to decompose the growth rate of USA from 1948 to 2010 and found that Solow residual, namely technological progress, accounts for more than half of the growth in the referred period (Jones and Vollrath, 2013). This result demonstrates the vital role played by technological progress in economic growth.

Cass (1965) and Koopmans (1965) built infinite horizon, general stochastic equilibrium models based on Ramsey (1928), in which the preferences of households endogenously determine the saving rate. However, the endogenization of the saving rate did not change the main result of the Solow-Swan model regarding long-run growth. These so-called “neoclassical growth models” also concluded that capital accumulation alone could not generate sustainable economic growth; the long-run growth rate is determined by the rates of population growth and technological progress.

Romer (1986) emphasized that Arrow’s (1962) inference is a consequence of the public good property of knowledge. Firstly, knowledge is a non-rival good. Once produced, different firms can use the knowledge simultaneously for different ends, without reducing its usability by other firms (Acemoglu, 2009). Secondly, knowledge is a non-excludable good; namely, the firm that produces the knowledge cannot prevent other firms from using it by patenting or keeping it secret, at least partially. Having these properties, the knowledge produced by a firm would spill over to the whole economy and create a positive externality that increases the productivity of all firms. Different from the previous studies, Romer (1986) treated knowledge as a kind of capital good and incorporated it into the production function as a separate production factor. Knowledge is produced by an R&D production function with diminishing returns property, similar to the capital goods production in other studies. The produced knowledge enhances not only the production possibilities of the firm by increasing its private knowledge stock but also those of all other firms by contributing to the aggregate knowledge stock of the economy. The economy-wide positive externalities more than compensate the decrease in the marginal product of knowledge for individual firms with the increase in their private knowledge stock. Therefore, knowledge exhibits increasing marginal product property, and the output production function exhibits increasing returns to scale in all factors of production. By this means, the continuous growth of knowledge stock enables long-term per capita output growth in the economy.

3 Endogenous Technological Change and Imperfect Competition

In the models reviewed so far, markets were assumed to be perfectly competitive, and firms’ incentives for obtaining market power were not explicitly taken into consideration. However, the last 200 years of the world economy shows that the extraordinary economic growth that occurred in the period is primarily due to innovations that arose as a result of intentional R&D activities of economic agents (Akcigit, 2017).

Firms’ R&D expenditures constitute a fixed cost and cause the emergence of increasing returns to scale property in production. Because of this property, firms’ average costs are always higher than their marginal costs. If firms charged a price equal to their marginal costs as in perfectly competitive markets, they could never cover the fixed costs of production and would suffer a loss. In this case, the firms would not have an incentive to make any innovation and to enter into the market in the first place. Therefore, increasing returns to scale requires imperfect competition, in which firms can charge prices that are higher than their marginal costs (Jones and Vollrath, 2013).

The AK models mentioned above maintained the perfect competition assumption and incorporated knowledge with public good properties in order to avoid the technical difficulties of modeling imperfect competition in the general equilibrium framework. After Spence (1976), Dixit and Stiglitz (1977) and Ethier (1982) developed a formal framework to deal with imperfect competition, “endogenous growth models” have emerged, which explain economic growth with the intentional R&D activities undertaken by firms in the hope of obtaining market power (Jones and Vollrath, 2013).

3.1 Product (horizontal) innovation and economic growth

The most important study that explains endogenous economic growth with product innovations is Romer (1990). There are four inputs to production in his model: capital, labor, human capital, and technology. Among these, human capital represents the rival, and technology represents the non-rival types of knowledge. Both human capital and technology can be accumulated indefinitely. When a new unit of technology, defined as a new product variety, is produced, the productivity of the whole economy increases. There are three sectors in the economy: final goods sector, intermediate goods sector, and R&D sector. Final goods sector produces a homogeneous final good using labor, capital, and human capital with a Dixit-Stiglitz (1977) type technology. Final goods can either be consumed by households or can be used by the intermediate goods sector in the production of capital goods. However, the intermediate goods sector also needs a new product design to make production. These designs are produced in the R&D sector using human capital and the aggregate knowledge stock of the economy and sold to the firms in the intermediate goods sector with an indefinite patent. The payments made for the patents constitute a fixed cost for intermediate goods production, and, hence, causes increasing returns to scale in this sector. Intermediate goods firms cover this fixed cost by charging a price above the marginal cost of production thanks to the market power they obtained due to the fact that intermediate goods varieties are not perfect substitutes. However, since entry is free, firms gain zero economic profit in the long run. Thus, the intermediate goods sector has a monopolistic competition structure, and technological progress arises as a result of the market incentives provided by monopoly rents.

On the other hand, since technology is assumed to be a non-rival and non-excludable type of knowledge, it spills over to all agents in the economy and increases the aggregate knowledge stock of the economy. The aggregate knowledge stock enters into the R&D production function linearly, meaning that every new product variety also enhances the productivity of human capital in the R&D sector. Therefore, the economy exhibits increasing returns to scale as a whole, which generates sustainable per capita output growth in the long run (Jones and Vollrath, 2013; Barro and Sala-i-Martin, 2004).

Grossman and Helpman (1991b) set forth a different interpretation of Romer (1990) study. Instead of a separate final goods sector producing a homogenous final good with a Dixit-Stiglitz type production function, Grossman and Helpman assumed that households have a Dixit-Stiglitz type utility function; namely, they gain utility directly from the variety of final goods. In this case, Romer’s (1990) intermediate good firms turn into monopolistically competitive final goods firms, which produce final goods varieties with the blueprints they bought from the R&D sector. Grossman and Helpman showed that their model reached the same results as Romer (1990). Rivera-Batiz and Romer (1991) developed another variety of Romer’s (1990) model, in which R&D production uses final products rather than human capital. Since R&D production requires just physical capital good investment, this model is referred to as the lab-equipment model (Acemoglu, 2009).

In all of the models mentioned above, there is a negative relationship between competition and economic growth, because of the need for market power to cover the fixed cost of innovations (Aghion, Akcigit, and Howitt, 2014). However, Bucci (2005) showed that if labor, the total amount of which is fixed, is used in all the three sectors of the economy as an input, an inverse-U type relationship emerges between competition and economic growth.

3.2 Process (vertical) innovation and economic growth

Another group of models explains the economic growth with process innovations. Process innovations can be defined as the introduction of vertically differentiated product varieties that provide more product services per unit cost, and that are perfect substitutes to the existing product varieties (Grossman and Helpman, 1991b). Process innovations can take the form of either decreasing production costs or increasing the quality of existing product varieties (Gopalakrishnan, Bierly, and Kessler, 1999; Acemoglu, 2009).

Since consumers have a clear preference for the new product varieties that are created by process innovations, new varieties render the old product varieties obsolete; therefore, firms that make the product innovations seize the market power of the producers of old varieties and attain monopoly rents. However, this market power is not for an indefinite period as for horizontally differentiated product varieties, and it will only last until another firm replaces them by creating a better variety with another process innovation. In this sense, models that explain economic growth with process innovations formalize the “creative destruction” idea of Schumpeter (1942), and, therefore, they are referred to as “Schumpeterian Growth Models”.

Aghion and Howitt (1992) introduce the seminal paper in this vein and present a “quality-ladder” model in which firms create higher-quality intermediate goods varieties with process innovations. There are three sectors in the economy. Final goods sector produces a homogeneous final goods using unskilled labor and intermediate goods with a constant-returns-to-scale technology. Intermediate goods varieties are used in the intermediate goods sector using skilled labor and product variety designs. Finally, the R&D sector produces higher quality intermediate goods designs randomly, which is represented with a Poisson process. The arrival rate of innovations is a positive function of the amount of skilled and specialized labor invested in the R&D sector. The knowledge created with innovations spills over to the economy and enhances productivity in the final goods sector. Entrepreneurs that make a successful innovation acquires an indefinite patent right on the product design and start production of the variety in the intermediate sector as a monopoly. Since new product varieties represent more efficient methods in production, they render the old varieties obsolete; therefore, the innovating firm replaces the old firms and acquires the market power until the next innovation. The higher the degree of market power that firms hope to acquire, the higher would be the average growth rate. In other words, competition and economic growth are negatively related. Grossman and Helpman (1991a) and Segerstrom, Anana, and Dinopoulos (1990) studies also developed similar models. Grossman and Helpman (1991b) introduced a different interpretation of Aghion and Howitt’s (1992) model, in which process innovations reduce the cost of production rather than increase the quality of the product, and showed that the same results would arise.

Aghion, Dewatripont, and Rey (1999) model the case in which the incentives of agents in the decision processes are different from those of the firm. According to them, intense competition reduces the slack in the market by making managers feel the pressure of failure and liquidation of the firm. Hence, managers become more eager to adopt new technologies and the growth rate increases.

Aghion, Harris, Howitt, and Vickers (2001) emphasize the tacit character of knowledge. If firms are required to discover the knowledge in innovations of their rivals by their own R&D efforts, then the standard leapfrogging assumption in the Schumpeterian approach must be replaced with a step-by-step technological progress assumption. In this case, for a firm to outrun the industry’s technology leader, it must first close the gap step-by-step by making the previous innovations of the leader with its own R&D efforts. However, once a firm catches the leader, not any patent can protect the leader from Bertrand competition. Since neck-to-neck competition renders life more difficult for firms in such industries, firms have a constant incentive to engage in R&D to surpass their rivals. Therefore, it is more likely to have a strong positive relationship between competition and growth in the industries where tacit knowledge is a restrictive barrier to imitation rather than the ones in which patent protection is the only barrier.

Aghion, Bloom, Blundell, Griffith, and Howitt (2002) investigate the effects of “leveled” and “unleveled” competition in the product markets. Two firms engage in Bertrand competition with differentiated products in industries where entry and exit are blocked. Innovations are assumed to be step-by-step as in Aghion et al. (2001) and enable firms to climb up the quality ladder. In this case, two opposite effects are in operation. First one is the “escape competition effect”, which describes the firms’ incentive to increase the distance with its rivals. The second one is the “Schumpeterian effect”, which refers to the firms’ incentives to appropriate the monopoly rents. In general, the escape competition effect dominates the Schumpeterian effect in low levels of competition, and vice versa in high levels of competition so that an inverse-U type relationship emerges between competition and economic growth.

3.3 Process and product innovations and economic growth