Working Papers in Economics

Transcription

1 University of Innsbruck Working Papers in Economics Foreign Direct Investment and European Integration in the 90 s Peter Egger and Michael Pfaffermayr 2002/2 Institute of Economic Theory, Economic Policy and Economic History Institute of Public Finance

2 Foreign Direct Investment and European Integration in the 90 s Peter Egger and Michael Pfaffermayr February 12, 2002 Abstract This paper examines the effects of the EU integration in the 90 s on the FDI relations in Europe. We find that the completion of the Single Market, the 1995 enlargement and the next enlargement are characterized by a substantial, positive anticipation effect in the period between the announcement and the formal establishment of each integration step. Neither the Single Market Programme nor the previous enlargement exhibited any further positive effects on intra-eu FDI volumes after their establishment. Key words: European Union; FDI; Gravity Models; Integration Effects JEL classification: F14; F15 Both authors: Department of Economic Theory, Economic Policy and Economic History, University of Innsbruck, Universitaetsstrasse 15, A-6020 Innsbruck, Austria, and Austrian Institute of Economic Research, Vienna, Po-Box 91, A-1103 Vienna. Peter.Egger@uibk.ac.at, and Michael.Pfaffermayr@uibk.ac.at 1

3 1 Introduction In the last decade, the debate on the trade effects of regionalism experienced a strong revival. While empirical research has overwhelmingly focused on the impact of preferential trade agreements on trade volumes (compare Frankel et al., 1998; Sapir, 2001; Soloaga & Winters, 2001), only a few attempts examine the importance of joining an integration group (i.e., regionalism) on foreign direct investment (FDI; see Brenton et al., 1999, for an exception). As far as preferential agreements affect impediments to both trade and multinational activity, one should expect an effect on bilateral FDI as well (compare Markusen & Venables, 1998, 2000). This paper looks at the tendency to new regionalism in Europe and concentrates on the impact of the completion of the Single Market Programme, the 1995 enlargement of the EU, and the Europe Agreements between the EU and the Eastern European accession countries on the evolvement of bilateral FDI in Europe. The empirical analysis follows the traditional literature on bilateral trade and FDI and is based on a gravity model. Our nested design of integration group and integration phase effects allows to isolate the impact of these EU integration steps. The next section presents the econometric set-up. Section 3 briefly describes the database and discusses the estimated integration effects, and section 4 concludes. 2

4 2 The Econometric Methodology We base the empirical analysis on a standard gravity model, which explains bilateral trade or FDI flows by relative and absolute factor endowments and transaction costs (Bergstrand, 1989, Brenton et al., 1999). In terms of the covariates, we stick the standard specification with the real GDPs and the popolation sizes of the home and the host country. log F DI ijt = β 1 log P OP it + β 2 log P OP jt + β 3 log(gdp/p OP ) it +β 4 log(gdp/p OP ) jt + α + π kp + µ ij + λ t + ε ijt, (1) where F DI ijt is the real FDI-stock held by country i in country j in year t. P OP denotes a country s population, (GDP/P OP ) is the the real gross domestic product per capita, which is a measure of relative endowments (capital labor ratios). α is the constant and ε the traditional remainder error term. In analysis of variance (ANOVA) terms, our specification is a nested design with two main effects, the bilateral ones (µ ij ) and the time effects (λ t ). The former control for all observed and unobserved, time-invariant influences like distance, common language, common borders, etc., while time effects capture all time-specific influences common to all bilateral relations, such as cycle effects. π kp are integration group specific integration phase effects, with the main effects nested therein. π is designed to vary across three integration phases (p = 1, 2, 3) and across bilateral integration group relations (k = 1,..., 20). For conceptual reasons and and to guarantee consistency of the remaining coefficients, all effects are treated as fixed. 1 1 We do not think about our sample of countries as randomly drawn from a large population of bilateral relationships (see Baltagi, 2001). 3

5 In our sample, Phase 1 covers the pre-single Market period The second phase refers to , where the Single European Market came into effect and, additionally, the Europe Agreements with the majority of the Central and Eastern European countries have been signed. Phase 3 starts in 1995, when Austria, Finland and Sweden have joined the EU. There are 5 types of home and host country groups involved. Group 1 refers to the EU12 members, group 2 contains the three joining members of 1995, group 3 comprises the remaining three EFTA countries (Iceland, Norway, Switzerland), group 4 are the accession countries (in our case: Bulgaria, Czech Republic, Hungary, Poland, Romania, Slovak Republic, and Slovenia), and group 5 is the rest of the world countries hosting OECD direct investment. The set of bilateral FDI relations between these integration groups comprises K = 4 5 = 20 elements, since FDI originating from the accession countries into other economies is negligible. These 20 bilateral integration group effects are observed in all three phases, which gives 60 integration dummies. The nested model structure in (1) requires an appropriate set of restrictions on the parameters π kp, µ ij, and λ t. Equivalently, one can omit all phase-group dummies referring to one particular phase (here Phase 3), and one particular bilateral integration group (here FDI from the non-european OECD countries into non-european economies) and one year effect and one bilateral effect, which is just a reparameterization of the problem (compare Baltagi, 2001). With such a design, the estimated integration effects have to be interpreted as deviations from the typical trade relationship between the rest of the included OECD economies in non-europe in phase 3. To come 4

6 up with estimates of the integration effects, we analyse the phase-to-phase differences of the estimated integration group integration phase effects and compare them across integration groups. These differemce-in-difference estimates have to be interpreted as the impact on bilateral FDI, which is only due to the institutional change between the integration phases. 3 Data, Regression Results and the Assessment of EU Integration Effects We use real bilateral stocks of outward FDI as the dependent variable, which is constructed as follows. We take the available book values of foreign assets (OECD, Foreign Direct Investment Statistics Yearbook) as an admittedly rough indicator of the nominal figures of outward stocks of FDI and convert them by country specific investment deflators and exchange rate indices to obtain proxies of real figures taking 1995 as the base year. A few observations are interpolated to obtain continuous time series. Real GDP and population are from IMF (International Financial Statistics). Altogether, the data set comprises 3642 non-missing observations, and it is unbalanced. - Table 1 - The regression results are by and large in line with theory. Bilateral outward FDI-stocks are the higher, the larger the home country is and the better it is endowed with capital as compared to labor (Egger & Pfaffermayr, 2000). We find that OECD economies predominantly have invested in small, relatively capital-rich countries, which squares with theory. The models of trade 5

7 and horizontal multinationals (Markusen & Venables, 2000) support the view that FDI flows into large markets, while the vertical view on FDI (Helpman, 1984) suggests that predominantly the relatively labor-rich countries host FDI. - Table 2 - In assessing the EU integration effects, we focus on different, time-invariant aggregates of bilateral FDI relations. We calculate phase-to-phase differences of the integration effects for various aggregates of integration groups and compare them with those of appropriate control groups. This represents difference-in-difference estimates of the integration effects. Table 2 displays the results. The first bloc of results refers to the impact of the Single Market Programme (SMP). One would expect a direct positive effect of the SMP on intra-eu12 FDI, in so far as it has directly increased the locational advantage (European Commission, 1996, p. 87). Its net effect is ambiguous, since the SMP reduces trade costs thereby favoring the concentration of production (exporting) versus multi-plant production (FDI). As Table 2 indicates, we cannot identify any further effect after the completion of the SMP. Intra-EU12 FDI has not grown faster after 1992 than FDI within other European country blocs (the rest of the European Economic Area) or between the EU12 and the rest of the world (RoW). Similar results are obtained when comparing intra-eu12 FDI with FDI from other OECD economies into the EU12 area (not reported). Hence, the SMP effects seem to have materialized already in the period between the Single European Act 6

8 of 1987 and the formal establishment of SMP in 1992 (compare European Commission, 1996, p. 95ff), i.e., our Phase 1. We find a clear anticipation of the 1995 enlargement effect between 1993 and 1995, but not earlier. Our estimates indicate that the real FDI volume between the three new member countries and the EU12 has grown by 100 (e ) = 26% faster than intra-eu12 FDI itself between this and the previous integration phase. Similarly, the increase in FDI was 24% stronger than the increase in FDI within the EU12 and within the new members together. Again, there is no additional effect in the period after the formal enlargement in The last phase ( ) already reveals a pronounced anticipation of the expected so-called Eastern Enlargement of the Union. According to our estimates, no immediate positive effect of the signment of the Europe Agreements ( ) or any anticipation of it has been experienced. Only the most recent integration phase is characterized by a strong increase in FDI of the EU15 into the seven CEEC (excluding the Baltic States), which is 38% stronger than the change of EU15 FDI into both the EFTA and non- European economies and about 31% stronger than intra-eu15 FDI itself. 4 Conclusions This paper assesses the impact of three different events in the EU integration process during the 90 s on bilateral European FDI relations: the Single Market Programme, the 1995 enlargement and the Europe Agreements between the EU and the CEEC. Previous research has identified a positive effect of 7

9 the announcement of the Single Market Programme as set out in the 1985 European Commission White Paper. We find that such anticipation effects are also common to other recent integration steps. Our difference-in-difference estimates suggest that anticipation effects on FDI typically take place between the announcement and the formal establishment of an integration event. On the one hand, FDI does not increase before the official announcement of an integration step. On the other hand, the integration effects seem to be exhausted with the formal completion. Hence, recent integration effects on FDI relations in Europe have been substantial but mainly anticipatory. 5 References Baltagi, Badi H. (2001), Econometric Analysis of Panel Data (Wiley: Chichester). Bergstrand, Jeffrey H. (1989), The Generalized Gravity Equation, Monopolistic Competition, and Factor Proportions Theory in International Trade, Review of Economics and Statistics 71, pp Brenton, Paul, Francesca Di Mauro, and Matthias Lücke (1999), Economic Integration and FDI: An Empirical Analysis of Foreign Investment in the EU and in Central and Eastern Europe, Empirica 26, pp Egger, Peter and Michael Pfaffermayr, Trade, Multinational Sales, and FDI 8

10 in a Three-Factors Model, Working Paper No. 0013, Universtity of Linz. European Commission (1996), Economic Evaluation of the Internal Market, European Economy, Reports and Studies No. 4. Frankel, Jeffrey A., Ernesto Stein, and Shang-Jin Wei (1998), Continental Trading Blocs: Are They Natural or Supernatural?, in Jeffrey A. Frankel (ed.), The Regionalization of the World Economy (University of Chicago Press: Chicago), pp Helpman, Elhanan E. (1984), A Simple Theory of International Trade With Multinational Corporations, Journal of Political Economy 92, pp Markusen, James R. and Anthony J. Venables (1998), Multinational Firms and the New Trade Theory, Journal of International Economics 46, pp Markusen, James R. and Anthony J. Venables (2000), The Theory of Endowment, Intra-Industry, and Multinational Trade, Journal of International Economics 52, pp Sapir, André (2001), Domino Effects in Western European Regional Trade, , European Journal of Political Economy 17, pp Soloaga, Isidro and Alan L. Winters (2001), Regionalism in the Nineties: 9

11 What Effect on Trade?, North American Journal of Economics and Finance 12, pp

12 Table 1 - FDI Gravity Regression Results Explanatory variables β t-statistic p-value Log of exporter population Log of importer population Log of exporter GDP per capita Log of importer GDP per capita Constant Observations 3642 R Estimated autocorrelation: ρ 0.62 F-statistic DF p-value F-tests Bilateral effects ; Time effects ; Group x Phase 1 effects ; Group x Phase 2 effects ; Bold p-values indicate significance at 5%.

13 Table 2 - Testing for European Integration Effects on Real Stocks of FDI Hypotheses Involved country groups Coefficient F-statistic P-value Single Market Programme versus Intra-EU12 vs. intra-eea rest Intra-EU12 vs. EU12 with RoW versus Intra-EU12 vs. intra-eea rest Intra-EU12 vs. EU12 with RoW EU 1995 Enlargement versus entrants with EU12 vs. Intra-EU entrants with EU12 vs. Intra-EU12 plus intra-entrants versus entrants with EU12 vs. Intra-EU entrants with EU12 vs. Intra-EU12 plus intra-entrants Europe Agreements (Eastern Enlargement) versus EU15 with applicants vs. EU15 with RoW EU15 with applicants vs. Intra-EU versus EU15 with applicants vs. EU15 with RoW EU15 with applicants vs. Intra-EU Test statistics are distributed as F(1, 3150). Bold p-values indicate significance at 5%.

14 Zuletzt erschienen Recent Discussion Papers 2002/01 Peter Egger and Michael Pfaffermayr: The Pure Effects of European Integration on Intra-EU Core and Periphery Trade. 2002/02 Peter Egger and Michael Pfaffermayr: Foreign Direct Investment and European Integration in the 90 s