Growth Theory and Growth Policy

Harald Hagemann, Stephan Seiter, Harald Hagemann, Stephan Seiter

• Publisher: Routledge
• Year: 2003
• ISBN:
  9781134510894
• Edition: 1
• Topic: Business
• Subtopic: Business General
• Index: Business

This collection examines the phenomenon of economic growth with admirable economic vigour and includes contributions from leading academic figures. Theoretical approaches, underpinned by original empirical work, will make this a book welcomed by students and academics of macroeconomics and growth theory.

1 Introduction

Harald Hagemann and Stephan Seiter

Since Adam Smith the analysis of economic growth was one of the main fields of interest in economic theory. Particularly from the late 1940s to the 1960s, many papers dealing with the explanation of growth were published. The controversy between Postkeynesian authors such as Harrod, Domar, Kaldor and Robinson and members of the neoclassical school such as Solow, Swan, Meade, Arrow or Phelps gave inspiring insights into the process of economic growth. The discussion focused on the question whether the long-run economic growth process would converge to a stable equilibrium or not. Besides many differences between these two schools of thought, most economists agreed, and still do, that technical progress is the main source of per capita income growth. For example, Solow (1957) gave evidence to this statement by applying the concept of growth accounting on the development in the USA in the period of 1909–49, and Kaldor (1957, 1961) stressed the key role of technological innovations for economic growth.

However, the overwhelming interest in growth theory declined from the early 1970s. The economic slowdown after the first oil price shock drew economists’ interest to the explanation of business cycles, which had been the dominant theme in the interwar period but, after a short boom of multiplier–accelerator models after the war, had almost fallen into oblivion when the historically unprecedented growth process of the 1950s and 1960s had shifted economists’ main research areas. The phenomenon of stagflation raised again the question whether monetary and fiscal policy should follow Keynesian theory or not. The period after 1973 differed from the post-war years not only with regard to the characteristics of the business cycle, but also with respect to the long-run development. The average growth rates of Gross Domestic Product as well as of productivity were significantly smaller after 1973 than before. The so-called productivity slowdown gave new impulses to growth theory, although first studies applied well-known approaches to explain this phenomenon (see, e.g. Bombach 1985).

In the 1980s, Paul Romer’s seminal paper (1986) could be seen as the starting point of a renaissance of growth theory that has received greater attention in economic science since then. The so-called New Growth Theory challenged the traditional neoclassical approaches because of the latter’s theoretical and empirical shortcomings. The Solovian-type models cannot explain endogenously steady-state per capita growth. Exogenous technical progress is the only source of productivity growth in the long-run equilibrium, i.e. per capita income will be constant, if technical progress does not accrue. Savings and investment decisions determine only the level of long-run productivity, but not its growth rate. Growth is independent of economic decisions. At least, the models do not cover the relations between economic activities and technical progress as well as productivity growth. The dominant source of growth comes from outside the model and is like ‘Manna from Heaven’. The main reason for this shortcoming is the assumption of diminishing marginal returns to capital. The incentive to invest and thereby income growth per capita will peter out as the capital–labour ratio increases. New Growth Theory in contrast offers endogenous explanations of economic growth by focusing on constant returns to capital caused by externalities.

Furthermore, the standard neoclassical growth models predicted convergence between economies when its assumptions are met. In general, economies with a low capital–labour ratio should be able to realise higher growth rates of per capita income than economies with a high capital–labour ratio do. If economies show identical parameters with regard to infrastructure, time preference, supply of labour, etc. they will reach the same per capita income in the long-run as long as technical progress is a public good that is internationally available. Under these circumstances, even the steady-state growth rate will be identical in all economies that have access to technological knowledge. Empirical studies gave evidence that convergence is only found in groups of countries that are closely linked via trade in goods, services and knowledge such as the Organisation for Economic Growth and Development (OECD) members do (see, e.g. Sala-I-Martin 1996). Divergence between different convergence clubs of countries seems to be more realistic than convergence.

The finding that growth theory did not cover the empirics of growth led for example Romer (1994) to discuss a list of stylised facts that characterise current economic growth. Competition between companies is seen as a decisive precondition for technical progress. In general, research and development (R&D) lead to new products and therefore to competitive advantages for firms. Successful innovators will gain market power that enables them to gain extra profits. This is the key incentive to invest in R&D and to bear the risk of failure. Furthermore, innovators have to accept that there is no rivalry in the use of knowledge. Competitors can make use of the newly gained knowledge at zero marginal cost. This externality enables them to challenge the position of the former innovator and/or to become innovators themselves. The fact that knowledge is a public good allows the reproduction of physical processes without increasing non-rival inputs. Consequently, neoclassical economists can still assume constant returns to the application of labour and capital that guarantees a consistent explanation of the income distribution. These stylised facts go beyond the ones Kaldor (1957) referred to in his early work on growth theory. With regard to the recent developments in growth theory, particularly three items on his list are of interest. He proclaimed that (1) productivity growth does not fall in the long-run, (2) the capital–labour ratio steadily increases and (3) differences in productivity growth are caused by variations in the investment ratio.

The chapters in Part I present different contributions explaining endogenous growth. Diagnosing a great separation of the modern literature on the theory of endogenous growth on the one hand and labour economics on the other hand, Martin Zagler deplores the lack of models which comprise mutual causalities of growth and employment. These, however, are very important to explain long-run economic development of modern capitalist economies. He therefore aims to close this theoretical gap by developing a two-sectoral model with a manufacturing sector characterised by monopolistic competition and the payment of efficiency wages thereby creating unemployment, and an R&D sector where all profits of the manufacturing sector are reinvested. Zagler also takes up Arthur Okun’s empirical observation for the US economy that there exists a rather stable macroeconomic relation between the decline in unemployment and the growth rate of real output. However, Zagler criticises the shortcomings of Okun’s analysis, namely a lack of causality in either direction and the static nature of Okun’s argument. Both deficiencies are overcome in his model, which develops a revised version of Okun’s law, which is dynamic in nature and implies a clear causality, which runs from unemployment to economic growth. Whereas Okun’s law is widely understood as a cyclical phenomenon (see Hagemann and Seiter 1999), Zagler’s reinterpretation clearly expresses a long-run relationship between the level (not the change) of unemployment and the rate of economic growth. His NAWRU, i.e. the non-accelerating wage rate of unemployment, on the other hand remains unaffected by changes in the growth rate. In other words, causality in Zagler’s model is not mutual but runs unidirectional from unemployment to economic growth. That measures to enhance economic growth will not exhibit any impact on the level of unemployment probably is that aspect of Zagler’s contribution which is most difficult to swallow for many growth economists.

In his chapter, Stephan Seiter analyses the key elements of new growth theory and the Kaldorian approach to explain endogenous growth. His contribution reveals that besides all differences between the new models of growth, the assumption of constant returns to any sort of capital is crucial with regard to the factors of production that are assumed to be the driving force of economic growth. Thus, the models of new growth theory can be interpreted as variations of the AK-model. Furthermore, it is shown that Kaldor’s contributions to growth theory are based on the same ideas as new growth theory is. Investment is characterised by externalities that improve the productivity of other inputs and future investment. Hence, investment is the engine of endogenous growth. However, new growth theory only focuses on the supply-side variables of growth. Any output that is produced will be sold, since entrepreneurs know the households’ preferences. Here, Seiter gives evidence that Kaldorian models offer more insights by including both the demand side of endogenous growth and the determinants of investment. The interplay of demand and supply may lead to a cumulative growth process with positive or negative growth rates. Finally, the author concludes that with regard to growth policy a one-eyed focus on the supply side is not sufficient to explain endogenous growth consistently.

Flora Bellone and Muriel Dal-Pont examine the important relationship between finance and growth in the modern literature on endogenous growth and contrast it with the empirics of the East Asian growth experience. In the literature on the latter monetary and financial factors were hardly placed in the centre of the analysis until the financial crisis of 1997. The authors state that in contrast to the impressive relationship between high saving and investment rates, i.e. the process of capital accumulation, and the strong economic growth found out for this region, the question whether and how finance does matter for growth is still a missing link. The authors investigate two distinct lines of the endogenous growth literature, the AK or linear approach and the neo-Schumpeterian approach, for their appropriateness to deal with the finance–growth relationship. Both types of models differ in their causal structure and highlight another transmission channel why finance matters for growth. Whereas the AK approach attaches the leading role to capital accumulation, the neo-Schumpeterian approach attaches this role to technological innovations. Bellone and Dal-Pont investigate the advantages and drawbacks of the two approaches in relation to the explanation of the long-run growth process in East Asia. One important observation is that the causality from finance to growth of total factor productivity is much stronger in later stages of development when innovations become an important source for growth than in earlier stages for which neo-Schumpeterian models therefore did perform rather badly until the Soft Budget Constraint approach has offered a much better understanding of the East Asian productivity puzzle. One important result of the authors’ survey of the literature is that different financial systems do better in promoting different economic activities and perform differently over time. Bellone and Dal-Pont therefore clearly express their preference for time-series analysis rather than for cross-country analysis and advocate a stages of growth theory, however not a simple one. The issue of the role of financial factors in the transition process from one stage of growth to the other is still a rather unresolved question. Moreover, even for the same stage(s) of growth great differences concerning the impact of financial structures need to be explained, which refers to the significant element of diversity in growth regimes.

Summing up Romer’s and Kaldor’s lists of stylised facts and the results of the chapters of Part I, it becomes evident that on the one hand investment, and most of all investment in knowledge, is an important pre-condition for growth. On the other hand, an appropriate mixture of competition and market power promotes investment and risk-bearing. The importance of investment for economic growth offers a link between short-run and long-run development. In the models of the new growth theory investment in different kinds of capital leads to external effects on the productivity of other companies. We must also take into consideration that investment does increase the production capacities and creates income in the capital goods-producing sector. Thus, changes in investment influence both potential growth due to externalities and economic fluctuations due to income effects. For this reason, growth of the potential output is path-dependent in the sense that the investment pattern of the past determines the current level of productivity as, e.g. in Arrow’s paper of 1962. A strict disintegration of cycle and trend is therefore too much a simplification.

Part II connects therefore two important streams in modern macroeconomic literature: the interaction between cyclical fluctuations and long-run economic growth. Equilibrium and Real Business-Cycle models, which flourished in the late 1970s and in the 1980s, normally abstracted from the long-run trend. Modern endogenous growth theory, which started to flourish in the mid-1980s, on the other hand, in most cases abstracted from cyclical fluctuations, as it had been the case in the earlier growth literature after Harrod from the mid-1950s to the early 1970s. However, with regard to reality an important link between the cycle and the trend, i.e. an element of path dependency, cannot be denied. No wonder that against the background of a complete separation of these two phenomena in most theoretical models Bob Solow in the early 1990s rightly stated: ‘A more urgent matter is the problem of medium-term macroeconomics and the interaction of growth and fluctuations’ (Solow 1991: 16). In the spirit of Solow’s statement Richard Arena and Alain Raybaut analyse two different views on learning, knowledge and cyclical growth: neo-Schumpeterian versus neo-Kaldorian approaches. Thus, as in Seiter’s article, the importance of Kaldor’s contributions, which is often neglected in the modern endogenous growth literature, is highlighted, here with a stronger emphasis on Kaldor’s analysis of cyclical fluctuations. Although there is a basic agreement with the leading neo-Schumpeterian proponents of new growth theory ‘that growth and cycles are related phenomena, with causation going in both directions’ (Aghion and Howitt 1998: 4), and with Kaldor who had complained quite often about the traditional dichotomy between these two phenomena, the authors nevertheless make clear some significant differences between neo-Schumpeterian and neo-Kaldorian approaches, for which growth equilibria are only benchmarks and the prevailing state is that of economic instability or cyclical growth. Furthermore, the authors reveal also some important differences between the neo-Schumpeterian approach and Schumpeter’s own view, who in assuming that learning is closely related to firms’ investment activities was closer to Kaldor than to some neo-Schumpeterians, from whom he also differed in his emphasis on the creation of bank credit as a necessary condition for economic development. However, Arena and Raybaut cannot deny that Aghion and Howitt successfully introduced important Schumpeterian ideas into the modern framework of endogenous growth theory, above all the decisive role of technological innovations for cyclical fluctuations as well as for long-run growth. Thus, the neo-Schumpeterian approach made a decisive contribution to overcome the traditional dichotomy between business-cycle theory and growth theory. However, rational expectations suggest that the authors’ showing that the mere introduction of a learning-by investing mechanism renders possible the coexistence of growth and cycles is based on a simple AK model with adjustment costs will not taste well for the great majority of both neo-Schumpeterian as well as neo-Kaldorian economists.

Christiane Clemens and Susanne Soretz focus in their chapter on the effects of individual-specific and economy-wide productivity shocks on intertemporal decision-making of risk-averse agents. This analysis is mainly based on the real business-cycle theory where technical progress leads to an increase in productivity. Consequently, the rate of return to investment increases, too. Agents will thereby change their savings behaviour as well as their labour supply. Uncertainty rises due to technical progress. Thus, households will save more to cope with the higher risk. However, different types of income are characterised by varying risks. Clemens and Soretz base their research on two standard models of new growth theory: the AK model and Romer’s pathbreaking model of 1986. In the former, capital is the only factor of production. Hence, profits are the only source of income, and agents have to bear the risk of uncertain capital income. In the latter model, labour is also an input in the production of final output. Thus, the income risk is additionally included in the model. In this setting income distribution influences economic growth, since (precautionary) savings depend on the individuals’ risk aversion. The authors show that low risk aversion may be harmful to economic growth, since individuals will save less than highly risk-averse agents do.

Olivier Bruno and Patrick Musso also deal with the integration of short-run and long-run phenomena. Again, households’ savings behaviour is the link between cycle and growth. While Clemens and Soretz refer to real business-cycle theory, Bruno and Musso stress the relevance of monetary policy and inflation for long-run economic development. The integration of monetary policy in an endogenous growth model is important because both phenomena are interrelated. On the one hand, in the second half of the 1990s, US monetary policy could be less restrictive since the high growth rates of productivity reduced the inflationary pressure due to tight labour markets. On the other hand, an increase in the interest rate caused by monetary policy can lead to a decline in investment and growth. To cover the relations between inflationary processes and households’ savings Bruno and Musso develop an overlapping generations model with endogenous growth. If, for example, the level of inflation or the volatility of inflation influence savings, then any monetary policy to stabilise the price level will affect the long-run growth rate. One can also expect that an increase in the supply of money raises inflation. Thereby individuals will expect a lower return to savings. Depending on the risk aversion of households, savings might increase. In an endogenous growth model, the long-run growth rate will increase, too. The authors’ analysis makes clear under which conditions the model will generate these results. Therefore, monetary policies that try to reach an inflation rate of zero could lower the growth potential of an economy. In this sense, money matters in the long-run. Furthermore, the volatility of inflation can also reduce economic growth if the return to savings depends negatively on the volatility of inflation. Again, it is the level of risk aversion that determines the final outcome.

The analysis of endogenous growth gives some interesting insights into the growth consequences of economic policies. A general conclusion that can be derived from the results in new growth theory is to stimulate investment, especially in knowledge and education. Since investment is characterised by positive externalities, a system of subsidies to internalise the externalities could make sense at first sight. The higher the externality of an investment is, the more it should be supported. Nevertheless, not all welfare effects are positive. Models based on Schumpeter’s analysis of the process of creative destruction make clear that the success of innovator B could destroy the market power of the former innovator A. Unfortunately, the actual scope of an externality is difficult to measure. Furthermore, welfare economics proved that the internalisation of externalities only makes sense when technological and not pecuniary external effects exist. Thus, policymakers face the well-known information problem.

The chapters in Part III deal with relations between government intervention and economic growth. Besides the difficulties linked with the concept of externalities we further have to keep in mind that the models of New Growth Theory abstract from many important variables. Normally, standard welfare theory does not focus on the behaviour and utility function of the social planner who as a benevolent dictator is expected to maximise social welfare. However, modern public choice theory has revealed that this person might have own goals shewants to reach. Ingrid Ott discusses howdifferent assumptions about the social planner’s behaviour affect endogenous growth. She could be either altruistic or egoistic. The latter means that she will maximise her utility by maximising the budget relative to total output in each period or by realising the maximum budget over her infinite planning period. Ott proves that these objectives are contrary, so the social planner has to decide which goal shewants to reach or howshe should weight them in her utility function. Furthermore, the consequences of her behaviour also depend on the type of input she provides and the way she finances the budget. The model shows that not only a benevolent dictator will choose the Pareto-optimum, but also an egoistic planner might realise the Pareto-optimal provision of public input. Therefore, the type of input, the social planner’s behaviour, and the way to finance determines whether social welfare is maximised or not.

In many growth models, human capital plays a crucial role for endogenous growth (e.g. Lucas 1988; Romer 1990). The efficiency of the education system is relevant for the rate of growth, since human capital formation has also internal and external effects: internal because education and training increase human capital and thereby productivity of the individual, external because the increase of human capital of individuals does also increase the average human capital of society as a whole. Thus, one possibility to foster growth is via education policy. Besides its effects on growth, such a policy is justified by its implications for income distribution. A better education increases the probability of achieving higher labour incomes. Education policywould then lead to a more even personal income distribution, as it had been the case in most advanced economies in the last century. Christiane Schäper develops an overlapping generations model based on the approach of Galor and Tsiddon (e.g. 1996) to reveal the effects of education policy on growth and distribution. The key elements are a threshold externality of technological progress and the diminishing complementarity between investment in capital and the influence of parents’ human capital on the children’s human capital accumulation. Simulations run by Schäper show that education policy can affect both income distribution and economic growth positively. As in Ott’s model, financing is decisive because of the incentives on investment in human capital. ...