Interdependent Growth in the EU: The Role of Trade by María García-Vega José A. Herce DOCUMENTO DE TRABAJO 2002-08

March 2002

* Universidad Complutense de Madrid, Spain. Email: mlgarcia@econ.ucsb.edu ** FEDEA and Universidad Complutense de Madrid, Spain. Mail Address: FEDEA. C/ Jorge Juan, 46. 28001 Madrid. Spain. Phone: +34-914350401. Email: herce@fedea.es

Abstract

After properly modelling growth externalities and using spatial econometric techniques we investigate whether economic integration promotes interdependent growth among countries. We conclude that this has been indeed the case for advanced OECD countries and that, for those countries belonging to the EU, through successive enlargements, the effect has been even stronger. More precisely, if every (trade) partner of a given country experiences an extra growth of 1 percentage point, this economy will profit from an extra 0.5 point, and if this country belongs to the EU it will have an additional increase of its rate of growth of 0.2 points. Both figures can be interpreted as growth externalities with the latter suggesting that an integration process like the one followed by the EU has an (positive) effect on growth.

JEL classification: F15, F43, O47, R11

Resumen

Tras modelizar las externalidades de crecimiento y utilizando técnicas de econometría espacial, investigamos en este trabajo si la integración económica promueve el crecimiento interdependiente entre países. Concluimos que este ha sido ciertamente el caso para los países vanzados de la OCDE y que, para los países de la UE, a través de sus sucesivas ampliaciones, el efecto ha sido incluso mayor. Mas precisamente, si cada socio (comercial) de un país determinado creciese un punto porcentual adicional, este último crecería 0,5 puntos adicionales que se incrementarían en otras 2 décimas de punto (0,2) de pertenecer a la UE. Ambas cifras pueden interpretarse como externalidades de crecimiento, pero la última sugiere que un proceso de integración económica como el seguido por la UE tiene efectos positivos sobre la actividad económica.

1. Introduction

Trade can promote growth in industrial countries by enhancing competition, promoting product differentiation, specialisation and economies of scale. Romer (1990), Grossman and Helpman (1991), Griliches (1995), Lichtenberg and van Pottelsberghe (1996), Coe and Helpman (1995), among others, have investigated thoroughly the effects of learning-by-doing, accumulation of human capital, spillovers and diffusion of knowledge. The prediction of some of those theoretical models is that trade can induce convergence of growth rates across countries when there are technological externalities such as international knowledge diffusion (Grossman and Helpman, 1991; Rivera-Batiz and Romer, 1991).

While there is a large literature on measuring productivity spillovers (Bernstein and Mohnen, 1998; Branstetter, 2001; Coe et al., 1997; Eaton and Kortum, 1996; Meyer, 2001), few studies have specifically examined the importance of trade-driven interdependent growth. In this paper we explore dynamic relations between OECD countries’ rates of growth and trade amongst them. In particular, we focus our attention on whether the EU process, through trade, has made the growth rates of its members more interdependent. Some studies before ours have studied the existence of links among different sectors of the economy. Miller and Spencer (1977) and Grinols (1984) use general equilibrium approaches to reflect interdependence, from an empirical point of view, De Bressons (1996) measures sectoral interdependence in firms. However these papers do not consider an explicit relationship among countries, by instance through trade. The paper that is closest in spirit to ours is Goicolea et al. (1998) which shows that trade among Spanish regions enhance their mutual growth. We adopt their approach to apply it to the EU integration process.

European economic integration provides an interesting case study in this regard. Garcia-Vega and Herce (2001) showed that EU countries have increased their trade and rearranged their trade patterns becoming more EU-focused over time. This fact can provide information about the nature and size of effects of the EU upon productive structures. In particular we found that EU countries have relatively increased their flows of trade with new and (geographically) closer EU partners. Has this had any influence on the interdependence of growth patterns in the EU? This question is important mainly for two reasons: first of all, with the launch of the euro the monetary policies are tied up for EU countries. As the debate on the policy implications of asymmetric shocks for monetary unions has illustrated, it is of the utmost interest to promote policies that enhance interdependence in countries forming these unions. Secondly, more interrelated markets mean bigger and freer markets for firms what may induce greater growth due to an increase in the demand, larger scale economies, etc.

¹ For a survey of the recent literature see Keller (2001)

Our approach in this paper consists in using spatial econometric techniques in order to examine how trade dependence between EU countries incites their mutual growth. As an empirical tool, spatial econometrics (Anselin, 1988) based on the multidirectional dependence present among observations in cross-sectional data sets, links very naturally with the idea that EU integration generates more economic relationships and therefore more trade, and how growth in any particular EU country affects growth in other country depending on the trade links between them. We show that there is evidence of this, and that the EU process has played a role making the rates of growth of its members more interrelated.

The rest of the paper is organised as follows. Section 2 contains a simple model of interdependent growth. The estimation of interdependent growth externalities using spatial econometrics is carried out in Section 3. This section contains the paper’s central result concerning trade, growth and economic integration. Finally, Section 4 summarises the results obtained.

2. Interdependence in a standard growth equation.

Our theoretical framework follows closely that of Garcia-Vega and Herce (2001). We consider the standard growth model. Suppose that there are J countries indexed , at any time ; this model states that output in country i at period , is produced using technology, , capital, and labor,

\[Y _ {i t} = A _ {i t} K _ {i t} ^ {\alpha} L _ {i t} ^ {\gamma}\tag{1}\]

We assume that interdependence among countries can be expressed as follows:

\[A _ {i t} = \eta_ {i t} \left(\prod_ {j = 1} ^ {J} Y _ {j t} ^ {\omega_ {i j t}}\right) ^ {\rho_ {i}}; \text { with } j \neq i\tag{2}\]

where reflects both technology for country i, through , and a “thick markets” externality, , that emerges as markets expand. This part of equation (2) thus implies that output in country i benefits from the fact that the market is growing, as well as bilateral relationships, as economic integration proceeds. This positive effect is due to the fact that markets are larger and more densely populated by firms and consumers and display a larger variety of goods and services and, as countries trade more, they become more interrelated what reinforces the previous effect. Also in equation (2), translates a growth externality that becomes larger the larger bilateral relationships are, measured through (trade, investment flows, distance, etc).

The externality set-up assumed in (2) has been proposed by Bertola (1992) where interdependence is established through private capital. It is also retaken in Goicolea et al. (1998) where output (market) in other regions or countries replaces capital as the explicit context where externalities emerge and bilateral relationships constitute the channel though which they flow.

Replacing (2) into (1) and taking logarithms one obtains:

\[y _ {i t} = \log \eta_ {i t} + \rho_ {i} \sum_ {j = 1} ^ {J} \omega_ {i j t} y _ {j t} + \alpha k _ {i t} + \gamma l _ {i t}; \text { with } j \neq i\tag{3}\]

where . Taking differences in (3) we get:

\[g _ {y i t} = c + \rho_ {i} \sum_ {j = 1} ^ {J} w _ {i j t} g _ {y i t} + \alpha g _ {k i t} + \gamma g _ {l i t} + \varepsilon_ {i t}; \text { with } j \neq i\tag{4}\]

where , with being a (constant) technological growth parameter and an error term, and that can be interpreted as a moving average of over the medium run adopted in order to simplify the procedure to obtain equation . This variable will be given a more precise role latter on.

Equation (4) applies to every country i. If we furthermore assume that ∀i we end up with a general formulation for an interdependent growth equation of the sort:

\[g _ {y} = c + \rho W g _ {y} + \alpha g _ {k} + \gamma g _ {l} + \varepsilon\tag{5}\]

where is a constant term ( )1× NJ column vector; , and are ( )1× NJ column vectors for output, capital and labour growth rates for all years and countries; _ρ , α and are parameters and is a block diagonal ( )NJ NJ × matrix with the following structure:

ωijt

ωij

² We could have assumed ωijt to be independent of time. This can be true in the short run but not in the longer run. This idea of a smooth process for ωijt relative to the process for output, at least, seems to us to be a convenient way to simplify the derivation of equation (4).

\[W = \left[ \begin{array}{c c c c} W _ {1} & 0 & \vdots & 0 \\ 0 & W _ {2} & \vdots & 0 \\ \vdots & \vdots & W _ {t} & \vdots \\ 0 & 0 & \vdots & W _ {N} \end{array} \right]\]

where each is a matrix of bilateral trade flows for that particular year whose elements are defined as follows:

\[w _ {i j t} = \left\{ \begin{array}{c l} \frac {x _ {i j t}}{\sum_ {s = 1} ^ {J} x _ {i s t}} & \text { if } i \neq j \\ 0 & \text { if } i = j \end{array} \right.\]

where x are exports from country i to country j. Note that the weights have been normalised (dividing by the total exports that country i has with the rest of the countries) in order to avoid biases due to the absolute size of trade flows or variation of their measurement unit across time. The arrangement of the W matrix implies that only contemporary spatial dependence is assumed (the elements out of the main diagonal are null).

There are basically two reasons for considering such specification. First, it makes explicit how relationships among countries generate a pattern of interdependent growth. Second, it permits the application of spatial econometric techniques to a specification based on a theoretical model. Spatial econometric techniques (Anselin, 1988), permit to study cross-section data with dependency relationships in space. This spatial dependence can be caused by measurement problems or because there exist complex dependence and interrelations in the studied phenomena based on “physical transfer of commodities, people, information”, “background geography” or “relates to more volatile levels of spatial regularity” (Haining, 1986). In particular, the parameter measures the importance of the “spatial lagged variable” (in spatial econometrics that makes reference to the variable that is the product of the weights matrix W with the dependent variable) or in our case the strength of the influence of trade dependence on growth. In other words, it indicates the extent of trade autocorrelation, after the other variables have been controlled for. Therefore, following this specification we test whether there is growth dependence among countries due to trade interactions.³

³ We explored different ways to construct these weights without significatively different results. Some of the alternatives where indicative that, below a certain level, trade does not influence growth patterns.

3. Interdependent growth among EU countries.

We estimate equation (5) for EU countries (fourteen countries, Belgium and Luxembourg have been added all through this paper) using labour and gross fixed capital investment data from 1970 to 1997 (388 observations) from the OECD “Statistical compendium”. Data for capital are obtained using the perpetual inventory method . A first step is to test whether there is spatial dependence for the data of trade and growth. This can be done using the scatterplot of Moran (Vayá et al., 2000) shown in figure 1. In this graph the dependent variable, is plotted in the x axis whereas the explanatory variable, , is plotted in the y axis. One can determine whether there is spatial dependence in the data and its structure is as follows. If the observations are dispersed in the four quadrants, there is not spatial dependence. On the contrary, if the set of points lies mainly in the upper-right and lower-left regions this indicates that there is a positive spatial correlation.

Figure 1 Moran scatterplot for EU countries: Growth in any one country against trade-weighted growth in the rest of the countries

From figure 1 one can observe that there is indeed positive spatial correlation. Consequently the next step is the estimation of the interdependent growth equation (5). However in our specification we can overestimate the parameter of the interdependent growth term if we do not control by other characteristics that are similar among EU countries. EU countries arguably can be in similar stage of development and they can be growing faster until their arrival to their technological frontier. In order to control by those convergence issues, we add to the regression the term , that is the logarithm of per capita GDP lagged one year. One would expect that the parameter has a negative sign indicating that, other things equal, richer countries grow less rapidly than poorer ones. Equation (5) can then be rewritten as:

δ

_t I

K K I _{t t t1} ( ) 1 δ ₊ = − +

⁴ Perpetual inventory method measures the stock of capital (Kt) in year t as follows: , where is the rate of depreciation of the capital (we assumed 5%; calculations done with 7% and 3% do not change the results) and is the gross fixed capital formation in period t. To obtain the value of the stock of capital at 1969, somewhat arbitrarily, we assume a capital to output ratio of 3 for every country.

\[g _ {y} = c + \rho W g _ {y} + \alpha g _ {k} + \gamma g _ {l} + \beta y _ {- 1} + \varepsilon\tag{6}\]

In estimating equation (6) the errors are not independent and therefore OLS estimation would be inconsistent. Rather, we perform maximum likelihood estimation as usually done in spatial econometric literature. The results of the estimation are reported in columns (iii) and (iv) of Table 1.

We find that the “growth externality” has a positive and statistically significant influence whether we include controls or not, confirming that there are trade-driven linkages across countries in the EU. This growth dependence among EU countries, enhanced by trade, enters with a coefficient of around 0.6. This suggests that if every trade partner of a given country experiences an extra growth of 1 percentage point, this country’s economy will grow around an extra 0.6 point as implied by the value of in the third column of table 1. While the main result in table 1 indicates that trade interactions affect the pattern of growth for countries of the EU, we cannot deduce an effect from the EU process itself. Of course, one can conclude that as the EU process has helped to increase trade amongst its members it has also promoted interdependent growth for a natural interpretation of the coefficient of the independent variable is that it increases with trade.⁵

In order to assess the influence of the EU process more precisely, we include dummy variables in the estimation to account for the fact that, along the period analysed, more and more countries have become members of the EU. We consider that growth interdependence among countries could happen in a reinforced way if, among other factors, any two countries belong to the same economic block, the EU in our case. This idea can be formally expressed modifying expression (2) as follows:

\[A _ {i t} = \eta_ {i t} \prod_ {R} \left(\prod_ {j = 1} ^ {J} Y _ {j t} ^ {\left(\omega_ {R, i j t}\right)}\right) ^ {\rho_ {R, i}}; \text { with } i \neq j \text { and } R = A l l, E U \text { and } E F T A\tag{2'}\]

where the different regimes of R refer to all countries, those belonging to the EU and those belonging to the EFTA⁶.

After replacing (2’) into (1) and following the same process leading to expression (5) we obtain:

\[g _ {y} = c + \rho_ {A l l} W _ {A l l} g _ {y} + \rho_ {E U} W _ {E U} g _ {y} + \rho_ {E F T A} W _ {E F T A} g _ {y} + \alpha g _ {k} + \gamma g _ {l} + \varepsilon\tag{5'}\]

where and are block diagonal matrices like W and where:

\[w _ {R, i j t} = \left\{ \begin{array}{l l} \frac {x _ {R , i j t}}{\sum_ {s = 1} ^ {J} x _ {R , i j t}} & \text { if } i \neq j \text { and both } i \text { and } j \text { belong to } R \text { at year } t \\ 0 & \text { otherwise } \end{array} \right.\]

The results of estimation of equation are reported in columns (iv) and (v) of table 1. The results indicate that belonging to the EU or EFTA increases the growth dependence of its members. This effect seems to be greater for EFTA countries. The coefficients are, however, not very significatively different from zero due to the fact that the variables and present high correlation amongst them.

⁵ We performed alternative estimations of equation (6) where W was split into various dummy matrices for different trade intensity intervals. Trade intensity was measured as in Goicolea et al. (1998). We found that below a certain threshold of trade intensity, the growth externality turned out to be non significant. This result is in line with the findings of Bottazzi and Peri, (2001). Bilateral trade data for up to 22 OECD countries from 1970 to 1997 where constructed within the context of the PhD. project of one of the authors and are available from her upon request (mlgarcia@econ.ucsb.edu)

⁶ EFTA: European Free Trade Association.

To solve that, we re-estimate equation (5’) including other non EU OECD countries: the United States, Canada, Japan, Norway, Switzerland, Australia and New Zealand. We have in this case 588 observations (21 countries and 28 years). The Moran scatterplot for this new set of countries (figure 2) shows that there is again spatial dependence for the data.

The results of re-estimating equation (5’) for this larger set of countries are shown in table 2. As in the previous estimation for EU countries, we find that trade relations promote interdependent growth. We find also evidence that belonging to a trade club or other type of union, the EU or EFTA, affects the growth of the country members. Our estimates suggest (column vi in table 2) that that if every trade partner of a given country experiences an extra growth of 1 percentage point, this country’s economy will grow around an extra 0.5 point as implied by the value of . Moreover, if this country belongs to the EU there is an additional 0.17 percentage point of growth added to the previous figure or an additional 0.16 point if it belongs to the EFTA. The collinearity problem we detected in the previous estimation has now almost disappeared. The overall growth externality seems now to be larger when more countries are added to the sample, with this increase due to a larger “club effect”. Without loosing significance, these results hardly change when we control by convergence issues⁷ (columns ii, viii, ix and x).

β

βUSA

⁷ The parameter β measures the convergence of the country to its technological frontier, whereas measures the catching-up with the US This last variable is defined as the difference between the logarithm of per capita GDP of any country with the logarithm of per capita GDP of the US.

In columns (ix) and (x) we added an inflation variable (GDP deflator) to the regressors in order to control for other unobserved characteristics of a club of countries, in particular that there has been co-ordination of macroeconomic policy amongst them, particularly in the case of the EU. The results show that growth externalities remain statistically significant and at about the same size previously found.

4. Conclusion

This paper uses bilateral trade data for EU and other OECD countries to measure trade related spillovers on growth. Several channels may be efficient to transmit spillovers, as our growth model shows, although we have opted for trade: namely, more trade interactions promotes interdependent growth. In our estimations we also allow for the effect of belonging to the EU, or any other club, to show up as a factor that reinforces this growth externality. Our empirical results show that an integration process like the one followed by the EU has an effect on the growth pattern of the country members. Although we have not strictly shown that integration (through trade) has increased the rate of growth in the EU, our results clearly indicate that deeper integration between countries leads to larger trade exchanges and to more interdependent growth patterns. As long as trade promotes interdependent growth, integration (through other policies) also promotes mutual growth.

The simple growth model presented in the paper captures the idea that integration and trade benefit the participating economies because the market at the reach of their firms increases. As an extension, it would be interesting to perform a more detailed analysis by sectors in order to see where these trade externalities concentrate and what has been the effect of the EU process on the productive structure of its members.

Furthermore, whereas this paper has focused only on trade as the element that makes economies interdependent, we think that the analysis of other aspects that are becoming increasingly important in the EU such as financial integration, the Euro or other EU-wide policies, as channels for the growth externalities of the sort analysed in this paper, can provide a deeper understanding about the linkages between integration and growth.

References

1. Aghion, P. and P. Howitt (1998) Endogenous growth theory, MIT Press.
2. Anselin, Luc (1988) Spatial Econometrics: Methods and Models. Kluwer Academic Publishers.
3. Baldwin, R. (1993), “On the measurement of dynamic effects of integration”, Empirica 20.
4. Bernstein, J. and P. Mohnen (1998) “International R&D Spillovers between U.S. and Japanese R&D Intensive sectors”, Journal of International Economics 44:315-38
5. Bertola, G. (1992) “Models of economic integration and localised growth”, Discussion Paper 651 CEPR.
6. Bottazzi, L. and G. Peri, (2001) “Innovation and Spillovers in Regions: vidence from European Patent Data” in www.econ.ucdavis.edu/faculty/gperi/Papers/euspill_final.pdf
7. Branstetter, L. (2001) “Are knowledge spillovers national or international in scope?. Microeconomic evidence from the U.S. and Japan”, Journal of International Economics 53:53-79.
8. Coe, D. and Helpman, E. (1995) “International R&D spillovers”, European Economic Review, 39:859-887.
9. Coe, D., E. Helpman, and A. Hoffmaister (1997) “North-South spillovers”, Economic Journal 107:134-149.
10. De Bresson, S. (1996) Economic interdependence and innovative activity: an input-output analysis. Aldershot, Edgar.
11. Eaton, J., and S. Kortum (1996) “Trade in ideas: patenting and productivity in the OECD”, Journal of International Economics 40:251-278.
12. Frankel, J.A. and D. Romer (1996) “Trade and growth: An empirical investigation”, NBER Working Paper 5476.
13. Garcia-Vega, M., and J. A. Herce (2002) “Integration and growth in the EU: the role of trade”, Información Comercial Española (Forthcoming).
14. Goicolea, A., J. A. Herce and J. J. de Lucio (1998) “Regional integration and growth: The Spanish case”, FEDEA DT 98-14.
15. Griliches, Z. (1995) “R&D and productivity: econometric results and measurement issues”, in P. Stoneman (ed.) Handbook of the Economics of Innovation and Technological Change, Blackwell, Oxford, pp.52-89.
16. Grinols, L. (1984) “A thorn in the lion’s paw: Has Britain paid too much for Common Market membership?”, Journal of International Economics, 16.
17. Grossman, G. and E. Helpman (1991) Innovation and growth in the global economy, Cambridge MA, MIT Press.
18. Haining, R. (1986) “Income diffusion and regional economics”, in Transformations through space and time, edited by Griffith and Haining Dordrecht. Martinus Nijhoff.
19. Helpman, E. and P. Krugman (1985) Market structure and foreign trade. Cambridge. MIT Press.
20. Keller, W. (2001) “International technology diffusion”, at: http://www.eco.utexas.edu/~keller/s110901.pdf
21. Lichtenbert, F., and B. van Pottelsberghe de la Potterie (1996) “International R&D spillovers: A re-examination”, NBER working paper #5688.
22. Meyer, J. (2001), “Technology diffusion, human capital and economic growth in developing countries”, UNCTAD working paper #154, Geneva, June.
23. Miller, M. and J. Spencer (1977) “The static economic effects of the UK joining the EEC: A general equilibrium approach”, Review of Economic Studies, 44.
24. Prewo, W. (1974) “Integration effects in the EEC”, European Economic Review, vol. 5.
25. Romer, P. (1990) “Endogenous technological change”, Journal of Political Economy 98:S71-S102.
26. Rivera-Batiz, L., and P. Romer (1991) “Economic integration and endogenous growth”, Quaterly Journal of Economics 106:531-555.
27. Vayá, E. and R. Moreno (2000), “La Utilidad de la Econometría Espacial en el ámbito de la ciencia regional”, FEDEA DT 2000-13.

RELACION DE DOCUMENTOS DE FEDEA

DOCUMENTOS DE TRABAJO

References

1. 2002-08: “Interdependent Growth in the EU: The Role of Trade”, María García-Vega y José A. Herce.

References

1. 2002-07: “Export Market Integration in the European Union”, Salvador Gil-Pareja y Simón Sosvilla-Rivero.

References

1. 2002-06: “Early Mortality Declines at the Dawn of Modern Growth”, Raouf Boucekkine, David de la Croix y Omar Licandro.

References

1. 2002-05: “Nearest-Neighbour Predictions in Foreign Exchange Markets”, Fernando Fernández-Rodríguez, Simón Sosvilla-Rivero y Julián Andrada-Félix.

References

1. 2002-04: “Demografía, empleo, salarios y pensiones”, Juan F. Jimeno.

References

1. 2002-03: “La reforma de la negociación colectiva en España”, Samuel Bentolila, Juan F. Jimeno.

References

1. 2002-02: “Efficiency Spillovers from Foreign Direct Investment in the EU Periphery: A comparative study of Greece, Ireland and Spain”, Salvador Barrios, Sophia Dimelis, Helen Louri y Eric Strobl.

References

1. 2002-01: “Non-Linear Forecasting Methods: Some Applications to the Analysis of Financial Series”, Oscar Bajo-Rubio, Simón Sosvilla-Rivero y Fernándo Fernández-Rodríguez.

References

1. 2001-23: “Age at first-time homeownership in Spain”, Namkee Ahn.

References

1. 2001-22: “Capital público y efectos desbordamiento. Un análisis del impacto de las infraestructuras sobre la actividad privada por Comunidades Autónomas”, Alicia Avilés Zugasti, Rosario Gómez García y José Sánchez Maldonado.

References

1. 2001-21: “Employment and public capital in Spain”, Xavier Raurich, Hector Sala y Valeri Sorolla.

References

1. 2001-20: “Son relevantes el capital humano y el mercado de trabajo en los modelos de contabilidad generacional?. Un estudio sobre el caso español”, Javier Alonso Meseguer.

References

1. 2001-19: “Duration of Fiscal Consolidations in the European Union”, Reyes Maroto Illera, Carlos Mulas-Granados.

References

1. 2001-18: “Car quality improvements and price indices in Spain”, Mario Izquierdo, Omar Licandro y Alberto Maydeu.

References

1. 2001-17: “Economic Integration and Regional Business Cycles: Evidence from the Iberian Regions”, Salvador Barrios y Juan José de Lucio.

References

1. 2001-16: “An Empirical Evaluation of Non-Linear Trading Rules”, Julián Andrada-Félix, Fernando Fernández-Rodríguez, María Dolores García-Artiles y Simón Sosvilla-Rivero. -Rivero.

TEXTOS EXPRESS

References

1. 2001-01: “La reforma de las pensiones en el contexto internacional”, José A. Herce y Juan F. Jimeno.

References

1. 2000-03: “Efectos sobre la inflación del redondeo en el paso a euros”, Mario Izquierdo y Simón Sosvilla-Rivero.

References

1. 2000-02: “El tipo de cambio Euro/Dolar. Encuesta de FEDEA sobre la evolución del Euro”, Simón Sosvilla-Rivero y José A. Herce.