Kaldorian approach to the economic growth of Greek regions

Kaldorian approach to the economic growth of Greek regions

Paschaloudis, D

The present paper reports an attempt to examine the tendencies of regional growth and convergence with respect to Greek regions. It does so by establishing the validity of Kaldor’s second Law of growth (or Verdoorn’s Law) and tests for regional convergence in the ‘conventional’ framework introduced by Barro and Sala-i-Martin. The empirical results suggest that regions of Greece converge at an extremely slow rate. This slow rate can be attributed to differences in regional specialisation. Moreover, is established that both the manufacturing and the service sector are subject to increasing returns. Therefore, the service sector should not be considered as a ‘passive’ sector, but rather as one of the ‘leading’ sectors of the Greek economy,

Keywords: Regional economic growth, Keynesian Neo-classical model, Regional convergence-divergence

I. Introduction

A major concern for economists in general, and for development economists in particular, is to find an explanation for the persistent disparities in the level of development between different regions within a country. Broadly speaking the problem at hand is ‘why levels of growth differ between regions?’ The question is simple; the answer less so. Traditionally, there are two opposing tendencies to this issue. First, the neo-classical approach based on the concept of general equilibrium and second the Post Keynesian approach, as formulated in the disequilibrating models of ‘cumulative causation’ or ‘centre and periphery,’ based mainly on the existence of increasing returns.

Ever since Lord Kaldor suggested that the statistical relationship between the rate of growth of labour productivity and that of output (also known as the Verdoorn’s Law) is an indication of substantial increasing returns to scale, the relation it has been surrounded by considerable debate.

Verdoorn’s Law is an indispensable element of Kaldor’s1 growth model, in which he highlights the importance of the manufacturing sector in the process of economic development. Many attempts have been made to test the validity of this Law. The empirical literature includes studies with respect to various countries and regions, both time series and cross section. In general, the majority of them provide empirical confirmation of this relation. These studies usually refer to the manufacturing sector and tend to take a narrow view by considering this sector as the only leading sector of the economy. With this in mind, it becomes evident that the service sector is typically ignored. The important point to emphasize is that modern economies can have more than one leading sector.

This paper has a contribution to make to the existing literature on regional convergence with respect to Greek regions. In particular, this paper has two purposes. The first is to shed some further light on whether or not there are increasing returns to scale in the manufacturing sector using regional panel data for Greece. McCombie and de Ridder (1984) came to the conclusion that the results from estimating Verdoorn’s Law using regional data provide a strong confirmation of the existence of substantial economies of scale in the manufacturing sector. But, in modern economies, the service or tertiary sector may also be considered as a ‘leading’ sector. This hypothesis is also tested in this paper using the Verdoorn relation. The second purpose of this paper is to confront alternative theories with empirical evidence regarding region convergence. In particular, the ‘conventional’ neo-classical approach to convergence, as outlined by Barro and Sala-i-Martin (1992a, 1995) is opposed empirically with the Post-Keynesian view.

The rest of this paper is organized in the following manner. Section II briefly develops the neo-classical and the Post-Keynesian approaches to regional economic growth and reproduces some of the criticisms that have been levelled against the former approach. The way that new approaches to regional growth map into contrasting ideas regarding the mechanisms of regional growth is then discussed. In particular, the main features of Endogenous Growth Theory and the New Economic Geography are also outlined briefly in this section. Section III specifies the mathematical formulation of Verdoorn’s Law together with the conventional neo-classical test for regional convergence. Section VI is then followed by a briefly discussion of the data used, while section V presents the econometric results. Finally, in the concluding section we offer a possible explanation for the results we obtain and suggest that might afford an interesting policy conclusion.

II. Why Levels of Growth Differ between Regions?

In this section we shall discuss some of the theories that have been put forward to explain regional growth and convergence. The models belonging to the neo-classical and Post-Keynesian tradition are outlined first, followed by a discussion of the more recent approaches by the Endogenous Growth Theory and New Economic Geography.

A. The Neo-Classical and the Post-Keynesian Approaches

In neo-classical theory, regional differences in levels of per capita income are supposed to converge in the long run, mainly through trade and factor mobility. The neo-classical approach depends critically upon the restrictive and unrealistic assumptions of perfect competition. We can note Richardson’s comments on this point:

“… The background assumptions of neo-classical growth theory are inapplicable to the regional economy …. Perfect competition cannot be assumed in regional economic analysis since space itself and the existence of transport costs limits competition; oligopoly, pure monopoly or monopolistic competition are much more appropriate market structures. Indeed, if we adopt neo-classical models in their pure unadulterated form there would be no such field as regional economics.” (Richardson 1973, p. 22) [Emphasis added].

Using the aggregate production function, which exhibits constant returns to scale, the neo-classical paradigm predicts that regional per capita incomes will finally converge. Convergence is brought about mainly through the free movement of factors of production. The argument runs as follows. Labour will migrate from poor regions to the prosperous regions, while capital will move from rich regions to poor regions. To the neo-classical economists the diffusion of innovations and technology is a meaningless topic. Technical progress, under the assumption of perfect competition, is available to all or at least, ‘plays a reinforcing role which accentuates the effects of investment and labour force growth’ (Borts and Stein 1964, p. 8).2 This suggests that the process of regional growth, in the neo-classical spirit is a static equilibrium process.

Nevertheless, several criticisms have been raised against the predictions, which this model has yielded because of various drawbacks underlying these predictions. First, the neo-classical approach is very supply orientated. This approach can be criticised on the grounds that it cannot provide a satisfactory answer to the question why the growth rates of capital, labour, and technical progress differ over time or between regions within a country. This is mainly because the sources of growth are treated as exogenous. Second, the use of the production function to explain growth performance is very aggregative. The essence of this aggregation is that the various sectors of the economy are assumed to have the same characteristics in production. This steams from the assumptions, which characterize perfect competition. But, as mentioned above, this assumption is unrealistic, when the primary concern is on regional economics. Third, the neo-classical approach makes the implicit assumption of constant returns to scale. This approach arises from the overriding emphasis on treating the economy as a whole. This assumption is also unrealistic. It is impossible to understand the growth and development process (and divisions between rich and poor regions within a country) without taking a sectoral approach. The sectoral approach involves a distinction between sectors with increasing returns on the one hand and sectors with diminishing returns activities on the other.

Kaldor, as an answer to the neo-classical approach, articulates three fundamental growth `laws,’ stressing the importance of industrialisation to the development process. In particular, Kaldor formulated them in a series of influential papers, published mostly in the 1970s.

These growth ‘laws’ are the following (Thirlwall 1983):

First Law: There exists a strong positive relation between the growth of manufacturing output and the growth of GDP.

Second Law: There is a strong positive relation between the rate of growth of productivity in manufacturing industry and the growth of manufacturing output.

Third Law: The faster the growth of manufacturing output, the faster the’ rate of labour transference from non-manufacturing to manufacturing, so that overall productivity growth is positively related to the growth of output and employment in manufacturing and negatively associated with the growth of employment outside manufacturing.

Kaldor’s first law stresses the importance of industrialisation in the development process. That is, the faster manufacturing output grows, the faster the GDP grows. What is the intuition behind this argument? There are two main explanations for this. The first is wherever industrial production and employment expand, labour resources are drawn from other sectors which have open or disguised unemployment, so that the labour transference to manufacturing does not cause a diminution in the output of these sectors and therefore productivity growth increases outside manufacturing (McCombie and Thirlwall 1994, p. 166).

The second explanation is rooted in the existence of increasing returns in the manufacturing sector, in the static and dynamic form. Before any progress can be made, it is necessary to understand exactly what is meant by static and dynamic returns to scale. Thirlwall emphasises this point.

“Static returns relate to the size and scale of production units and are a characteristic largely of manufacturing where in the process of doubling the linear dimensions of equipment, the surface increases by the square and the volume by the cube. Dynamic economies refer to increasing returns brought about by ‘induced’ technical progress, learning by doing, external economies in production, and so on. Because economies of scale result from increased product differentiation, new processes, new subsidiary industries, and so on, it was Young’s contention that they cannot be discerned adequately by observing the effects of variations in the size of an individual firm or of a particular industry. Economies of scale and increasing returns derive from general industrial expansion, which should be seen as an interrelated whole or as an interaction between activities.” [emphasis in the original] (Thirlwall 1983, p. 349).

Furthermore, economies of scale may be either internal or external. Internal economies may arise within firms on account of engineering or managerial considerations; external economies may be obtained by the geographical concentration of activities, so that services and other facilities may be shared in common (Chisholm 1990, p. 72).

All of these considerations are relevant for the way in which we view regions and their economic growth. Perhaps the main point of relevance is the following. Kaldor’s second growth law (Verdoorn’s Law) deals with the positive relation between the rate of growth of productivity in manufacturing industry and the growth of manufacturing output. In the present context, two features are of particular relevance. First, Verdoorn’s Law indicates the industrial sector as the leading sector of the economy or ‘the engine of growth.’ This argument arises from the overriding emphasis on the sectoral approach to the growth process. The manufacturing sector is subject to a faster growth of productivity, whilst the remaining sectors are ‘passive’ responding to the growth of the former. This, according to Kaldor, indicates the dual nature of modern economies. But beyond this, the second important feature of the Verdoorn’s Law is that it provides a base or a fundamental element of the cumulative causation models. Indeed, through the Verdoorn effect, the industrial growth can result in a cumulative process through the concentration of economic activities. In the context of regional economies, scale economies are of crucial importance. Due to the existence of increasing returns to scale, captured in the Verdoorn relation, growth rates not only differ among regions, but are also perpetuated. Internal and external economies of scale result in a geographical concentration of economic activities. This fact intersects with the Keynesian emphasis on disequilibrium processes, giving scale economies considerable status as an important mechanism, allied with multipliers and the Verdoorn effect, creating a cumulative growth processes.

The Verdoorn effect, therefore, provides a mechanism whereby a region, which grows faster than other regions will obtain productivity, or quality advantage, which will probably mean that it will continue to grow more rapidly, reinforcing the disequilibrium effects (Chisholm 1990). Kaldor recognised the importance of increasing returns to scale in determining regional growth rates. Kaldor offers a broader definition of increasing returns and the function of Verdoorn law in the context of cumulative causation models. A quotation from Kaldor gives a clear picture of his ideas:

“[Increasing returns to scale] are not just the economies of large-scale production, commonly considered, but the cumulative advantages accruing from the growth of industry itself – the development of skill and know-how; the opportunities for easy communication of ideas and experience; the opportunity of ever-increasing differentiation of processes and of specialisation in human activities. I…] One aspect of this [the Verdoorn’s Law] is that as communication between different regions becomes more intensified (with improvements in transport and in marketing organisation), the region that is initially more developed industrially may gain from the progressive opening of trade at the expense of the less developed region whose development will be inhibited by it.” (Kaldor 1970, pp. 315-6).

B. Endogenous Growth and New Economic Geography

In recent years, doubts have crept in the validity of the traditional neoclassical and the Post-Keynesian model. The 1980s and 1990s have seen the earlier growth models augmented by a new generation of growth theories. Research on economic growth has diverted into a new context provided by an entirely new strand: the New or Endogenous Growth Theory. According to Armstrong and Taylor (2001), this has given a welcome fillip to what had become a rather sterile debate between supporters of neo-classical and post-Keynesian views of regional growth.

Endogenous growth theory is build upon the premise that long-run growth in per capita output depends on investment decisions, rather than unexplained or exogenous improvements in technology, which are generated within the growth process. Thus, endogenous growth theory moves towards an explanation for the causes of technological progress.

This concept of technological progress has led to the development of endogenous growth models, which put strong emphasis on the dictum that technology is created intentionally. Following the spirit of this approach, progress in technology is the outcome of human capital formation and activities relating to the research and education and is incorporated into capital, defined in ‘broad’ terms. From this perspective endogenous growth models with ‘broad’ capital may deem as an extension of the neo-classical model. However, in the context of endogenous broad capital technological progress is no longer determined exogenously but it is an inherent element of the mechanisms of the economic system. In particular, the endogenous broad capital model modifies the traditional neo-classical production function in such a way to include externalities to investment.

The notion of externalities, especially in Arrow’s sense3, is the cornerstone of the endogenous growth theory. Romer (1986) claims that externalities due to investment in capital stock are generated through ‘knowledge spillovers’ and ‘learning by doing’ effects. It these externalities that make technology a ‘public’ good and, thus, technological progress is generated within the system as an outcome of the growth process. Romer developed the most instructive model in this category in series of influential papers, published mainly in the late 1980s and early 1990s.4 Romer argues that technology is a non-rival input; that is, its exploitation by one firm does not rule out its exploitation by another. In his formal model the important feature is the distinction between the rival component of knowledge and the non-rival technological component. Hence, the treatment of knowledge as a non-rival good provides an explanation for knowledge spillovers. A central tenet of Romer’s thesis is that firms can patent technological inventions and innovations, which gives them the exclusive right to produce new goods. In turn, these innovations create new ‘general’ knowledge, which is freely available to all firms. Romer’s (1990a) model explicitly describes the contribution of the research sector, which generates endogenous innovations.

Nevertheless, the best application of the notion of externalities can be found in the recent models that explain the rise and success of new industrial districts5 and to the models of New Economic Geography.

Krugman’s view of economic geography is, as Martin and Sunley (1996) point out, non-equilibrium one. In Krugman’s world, an automatic process of spatial equilibrium with equal factor payments does not exist. The uneven balance across space is attributed to similar mechanisms as in Kaldor’s model, namely the operation of dynamic increasing returns, which provide comparative advantages to a selected number of localities. Krugman argues that once an initial regional advantage is established it may become cumulative through backward and forward linkages, which ‘lock iri regions to a given growth pattern. Krugman does not deny the relevance of its theory to Kaldorian approach and acknowledged this by mentioning that

This clear dependence on history is the most convincing evidence available that we live in an economy closer to Kaldor’s vision of a dynamic world driven by cumulative processes than the standard constant returns model.” (Krugman 1991, pp. 9-10).

III. The Model

The primary purpose of this section is to articulate the mathematical framework upon which the empirical analysis will be conducted. In particular, the specification of Verdoorn’s law and the neo-classical expression for regional convergence are discussed.

A. Specification of Verdoorn’s Law

VI. Conclusions

The validity of Verdoorn’s Law at the regional level of analysis has confirmed for a variety of case studies. For example, recent studies established empirically the validity of Verdoorn’s Law in the regions of USA (Bernat 1992), the Spain regions (Leon-Ledesma 2000) and the regions of Japan (Casetti and Tanaka 1992). Despite the important implications for regional growth, Verdoorn’s Law and the Kaldorian approach, in general, did not receive much attention with respect to Greek regions. The existing empirical literature on Greek regions relies exclusively on the neo-classical approach. This paper adds to the existing literature by estimating Verdoorn’s Law using data referring to the regions of Greece.

Previous sections have documented the validity of the Verdoorn’s Law across the regions of Greece. The results provide considerable support to the thesis that Verdoorn’s Law holds in the case of Greek regions. Indeed, there is overwhelming support for the hypothesis of increasing returns to scale in the manufacturing sector. However, the manufacturing sector cannot be considered as the only leading sector of the economy. As the results indicate, the service or tertiary sector is also subject to considerable increasing returns to scale while this is not the case for agricultural sector. In the long run, which means considering a regional economy over a period of time as opposed to a point in time, some regions of Greece were able to grow due to a shift from agriculture to tertiary activities. During the last ten years Greece has been able to incorporate advanced technology in the service sector. This fact may explain, to a certain extent, the increasing returns in this sector. The fundamental proposition seems to be that the service sector cannot be considered simply as a passive sector, but rather as one of the leading sectors of the Greek economy. The large degree of returns to scale can also be considered as the result of economy- wide technical progress.

Another interesting conclusion refers to the tendencies towards regional convergence or divergence. To the extent that the determinants of Verdoorn coefficient vary between sectors, this coefficient may also vary between regions depending on their sectoral specialisation. The established variation of Verdoorn’s coefficient between activities and regions indicates either that Greek regions diverge or converge to a slow rate. The slow process of convergence is also confirmed by the neo-classical approach using extended data sets. From a sectoral point of view, the slow rate of convergence may be attributed to differences in sectoral specialisation across regions. Regions specialised in traditional activities are fall behind from regions that specialised in sectors, which are subject to increasing returns or with a high Verdoorn coefficient.

One final point must be addressed explicitly. Kaldor has argued in many of his writings that it is impossible to understand the growth and development process without taking a sectoral approach, distinguishing between increasing returns activities on the one hand and diminishing returns activities on the other. As a consequence, the main thrust of economic policy as far as regional economic development is concerned should be aimed at encouraging private and public investment (mainly infrastructure) in sectors that are subject to increasing returns to scale or in sectors with higher Verdoorn coefficients attached to them. Furthermore, provision of regional infrastructure should be oriented not only to the ‘traditional’ activities, such as transportation and communication systems, but also to the development of educational and research institutes in poor regions, in order to enhance indigenous human capital and local talents.

(Received 10 April 2001: Revised 14 November 2002)

1 In particular, see Kaldor (1966, 1970).

2 In other words, technology is treated as a public good.

3 Kaldor (1957) and Arrow (1962) are considered to be the ‘progenitors’ of endogenous growth theory. See McCombie and Thirlwall (1997).

4 See Romer (1986, 1990a, 1990b, 1993, and 1994).

5 This approach usually pays attention to specific regional case studies, which allow for a detailed analysis of the complex interacting forces that determine the evolution of a local system, by combining economic, social and cultural factors. See for example Harrison (1992), Storper (1992), Markusen (1996), and Acs (2000).


Acs, Z. (Ed.) Regional Innovation, Knowledge and Global Change, Pinter, 2000.

Alexiadia, S. Verdoorn’s Law and the Process of Cumulative Causation: The Case of Greece. Unpublished MA Dissertation University of Kent. 1998.

Armstrong, H., and Taylor, J. (Eds.) The Economics of Regional Policy. London, 2001.

Arrow, K. “The Economic Implications of Learning by Doing.” Review of Economic Studies 29 (1962): 155-73.

Barro, R., and Sala-i-Martin, X. “Convergence.” Journal of Political Economy 100 (1992a): 223-51.

“Regional Growth and Migration: A Japanese-US Comparison.” Journal of the Japanese and the International Economics 6 (1992b): 256-78.

Economic Growth. 2nd Edition, Mit Press, 1995. “Technology Diffusion, Convergence and Growth.” Journal of Economic Growth 2 (1997): 1-25.

Baumol, W. J. “Productivity Growth. Convergence and Welfare: What the Long-Run Data Show.” American Economic Review 76 (1986): 1072-85.

Bernat, G. A. “Does Manufacturing Matter? A Spatial Econometric View of Kaldor’s Laws’.” Journal of Regional Science 36 (1996): 463-77.

Borts, G. H., and Stein J. L. Economic Growth in a Free Market. New York, Columbia University Press, 1964.

Casetti, E., and Tanaka, K. “The Spatial Dynamics of Japanese Manufacturing Productivity: An Empirical Analysis by

Expanded Verdoorn Equations.” Papers of Regional Science Association 71 (1992): 1-13.

Chisholm, M. Regions in Recession and Resurgence. London: Allen and Unwin, 1990.

Dixon, and Thirlwall. “A Model of Regional Growth Rate Differences on Kaldorian Lines.” Oxford Economic Papers 27, 1975. Drakopoulos, S., and Theodossiou, I. “Kaldorian Approach to Greek

Economy.” Applied Economics, December, 1991.

Fingelton, B., and McCombie, J. S. L. “Increasing Returns and Economic Growth: Some Evidence from the European Union Regions.” Oxford Economic Papers 50 (1998): 89-105.

Harrison, B. “Industrial Districts: Old Wine in New Bottles?” Regional Studies 26 (1992): 469-83.

Kaldor “A Model of Economic Growth.” Economic Journal 67, 1957. Kaldor, N. Causes of the Slow Rate of Economic Growth of the United Kingdom. Cambridge University Press, 1966.

. “The Case for Regional Policies.” 1970. In Targetti, and Thirlwall (Eds.), The Essential Kaldor. London: Duckworth, 1990.

Krugman, P. “Increasing Returns and Economic Geography.” Journal of Political Economy 99 (1991): 82-499.

. Development, Geography and Economic Theory. MIT Press, Cambridge, MA, 1995.

Leon-Ledesma, M. “Economic Growth and Verdoorn’s Law in the Spanish Regions, 1962-91.” International Review of Applied Economics (No. 1 2000): 55-69.

Markusen, A. “Sticky Places in Slippery Space: A Typology of Industrial Districts.” Economic Geography (1996): 293-313. Martin, R., and Sunley, P. “Paul Krugman’s Geographical Econom

ics’ and Its Implications for Regional Development Theory: A Critical Assessment.” Economic Geography 72 (1996): 259-92. . “Slow Convergence? New Endogenous Growth Theory

and Regional Development.” Economic Geography 74 (1998): 201-27.

McCombie, J. “Kaldor’s Laws in Retrospect.” Journal of Post Keynesian Economics 5 (No. 3 1983): 414-29.

McCombie, J., and deRidder, J. ‘The Verdoorn’s Law Controversy: Some New Evidence Using US Data.” Oxford economic papers, June, 1984.

McCombie, J., and Thirlwall, A. P. Economic Growth and Balance of

Payment Constraint. London: St. Martins Press, 1994.

“The Dynamic Harrod Foreign Trade Multiplier and the Demand-Oriented Approach to Economic Growth: An Evaluation.” International Review of Applied Economics 11 (1997): 5-26.

Novell, J. P., and Marshal, E. V. “Kaldor’s Laws and Spatial Dependence: Evidence for the European Regions.” Regional Studies 33 (1999): 443-51.

Romer, D, Advanced Macroeconomics. McGraw-Hill, 1996.

Romer, P. “Increasing Returns and Long Run Growth.” Journal of Political Economy 94 (1986): 1002-37.

_. “Growth Based on Increasing Returns due to Specialisation.” American Economic Review 77 (1987): 56-62.

. “Endogenous Technological Change.” Journal of Political Economy 98 (1990a): S71-102.

. “Are Non-Convexities Important for Understanding Growth?” American Economic Review 80 (1990b): 97-103.

. “Ideas Gaps and Object Gaps in Economic Development.” Journal of Monetary Economics 32 (1993): 543-73.

,. “The Origins of Endogenous Growth.” Journal of Economic Perspectives 8 (1994): 3-22.

Richardson, H. Regional Growth Theory. Macmillan, 1973.

Siriopoulos C., and Asteriou, D. “Testing for Convergence Across the Greek Regions.” Regional Studies 32 (1998): 537-46.

Storper, M. “The Limits to Globalisation: Technology Districts and International Trade.” Economic Geography 62 (1992): 60-93. Targetti, T., and Thirlwall, A. P. (Eds.) The Essential Kaldor. London: Duckworth, 1990.

Thirlwall, A. P. “A Plain Man’s Guide to Kaldor’s Growth Laws.” Journal of Post Keynesian Economics, Spring, 1983.

Verdoorn, P. “Fattori Che Regolano lo Sviluppo Della Productivita del LA)voro.” LIndustria. Translated by A. P. Thirlwall, 1949.

D. Paschaloudis and S. Alexiadis*

*Professor. Technological Educational Institute of Serres, Terma Magnisias, 62 104 Serre, Hellas, Greece, (Tel) +302321-49121, (Fax) +302321-49135, (E-mail) dpasch@telser.gr; Part time lecturer, Technological Educational Institute of Serres, Greece, respectively. We are grateful to three anonymous referees for critical comments and helpful suggestions for improvement on an earlier draft of the paper. Any errors or omissions remain our responsibility.

Copyright Institute of Economic Research Winter 2001

Provided by ProQuest Information and Learning Company. All rights Reserved