![Sample our Global Development journals, sign in here to start your access, latest two full volumes FREE to you for 14 days]({"type":"image" , "src" : "/sda/5929/TOC-Subject-Sample-2021-Global-Development.jpg"})
ABSTRACT
The regional impact of European Union association agreements: an event-analysis approach to the case of Central and Eastern Europe. Regional Studies. Although European Union association agreements are generally seen as welfare improving, relatively little is known about their spatial–distributional consequences. Drawing on the pre-accession experience of Central and Eastern Europe (CEE), this paper estimates the regional growth effects of such agreements using an event-analysis approach. It finds a positive growth effect which is, however, not constant across association phases or types of regions (e.g., specialized regions gained most in the early association period and service-oriented economies gained most near accession); and it discusses the implications of this in relation to debates on regional growth in the CEE and for future waves of association.
摘要
欧盟联合协议的区域影响：中欧及东欧案例的事件分析方法，区域研究。儘管欧盟联合协议一般被视为能够增进福祉，但相对而言，对其空间分佈结果却鲜为人知。本文根据中欧及东欧（CEE）正式加入之前的经验，运用事件分析方法来评估这些协议的区域成长效应。但本文发现，正成长效应在各联合阶段或区域类别之间并不一致（例如专殊化的区域在联合初期获得最多，而服务导向的经济在接近正式入会时获得最多）；本文接着探讨上述发现之于CEE区域成长的辩论和未来的联合风潮之意涵。
RÉSUMÉ
L’impact régional des accords d’association de l’Ue: une analyse d’événement vis-à-vis de l’Europe centrale et orientale. Regional Studies. Bien que de l’avis général les accords d’association de l’Ue améliorent le bien-être, on sait relativement très peu de leurs effets spatiaux et redistributifs. Puisant dans la préadhésion des pays de l’Europe centrale et orientale (Peco), cet article cherche à évaluer les effets sur la croissance régionale de tels accords à partir d’une analyse d’événement. Il s’avère un effet positif sur la croissance qui n’est cependant constant ni suivant les phases d’association, ni en fonction de la typolgie des régions (p.e. les régions spécialisées ont profité le plus pendant les premières phases d’association et les économies axées sur les services ont tiré le meilleur profit au moment de l’adhésion); et on en discute les conséquences dans le cadre des débats à propos de la croissance régionale aux Peco et des vagues d’association futures.
ZUSAMMENFASSUNG
Regionale Auswirkung der Assoziierungsabkommen der EU: ein Ereignisanalysen-Ansatz für den Fall von Mittel- und Osteuropa. Regional Studies. Obwohl die Assoziierungsabkommen der Europäischen Union generell als wohlstandsverbessernd gelten, ist relativ wenig über ihre Konsequenzen für die räumliche Verteilung bekannt. In diesem Beitrag untersuchen wir auf der Grundlage der Erfahrungen in Mittel- und Osteuropa vor der Assoziierung die Auswirkungen solcher Abkommen auf das regionale Wachstum mithilfe eines Ereignisanalysen-Ansatzes. Wir stellen einen positiven Wachstumseffekt fest, der jedoch in den verschiedenen Assoziierungsphasen oder Regionstypen nicht konstant ist (z. B. profitierten spezialisierte Regionen in der frühen Assoziierungsphase und dienstleistungsorientierte Ökonomien kurz vor der Assoziierung am stärksten), und erörtern die Auswirkungen dieses Effekts auf die Debatten über regionales Wachstum in Mittel- und Osteuropa sowie auf künftige Assoziierungswellen.
RESUMEN
Impacto regional de los acuerdos de adhesión de la UE: un enfoque de análisis de eventos para el caso de Europa central y oriental. Regional Studies. Aunque en general se considera que los acuerdos de adhesión de la Unión Europea son una mejora para el bienestar, se sabe relativamente poco sobre sus consecuencias para la distribución espacial. A partir de la experiencia de preadhesión de Europa central y oriental (ECO), en este artículo calculamos los efectos de tales acuerdos para el crecimiento regional mediante un enfoque de análisis de eventos. Observamos un efecto de crecimiento positivo que, sin embargo, no es constante en todas las fases de adhesión o todos los tipos de regiones (p. ej., las regiones especializadas avanzaron más en el periodo de adhesión temprana y las economías orientadas en los servicios avanzaron en el periodo más cerca del momento de adhesión). Analizamos las repercusiones con relación a los debates sobre el crecimiento regional en la ECO y para las futuras olas de adhesión.
ACKNOWLEDGEMENTS
The authors are extremely grateful to Ivan Turok for his speedy and diligent handling of the paper, as well as to him and the anonymous referees for their constructive comments to the earlier version of the manuscript. The authors also benefited from comments made by members of the Economic Geography Specialty Group of the Association of American Geographers.
SUPPLEMENTAL DATA
Supplemental data for this article can be accessed at http://dx.doi.org/10.1080/00343404.2016.1198472
Notes
1. For ex ante evaluations of EU integration on regional growth under a general equilibrium framework, see, inter alia, d’Artis (Citation2001), Bradley (Citation2006), and Varga and Baypinar (Citation2016).
2. Assuming that ‘selection’ into an association agreement is positively related to a country's growth performance.
3. The inclusion of these country-specific trends essentially expresses regional growth (the dependent variable) as a deviation from each region's national (country-specific) growth trajectory. As the signing of EU agreements can safely be assumed to be exogenous to such deviations, the shift coefficients arguably capture the causal effect of these agreements.
4. In the descriptive analysis a regional-level measure of export orientation is also used (calculated using EUROSTAT data on national-level exports by manufacturing sector and data on each region's contribution to national employment in each sector). The method of construction of this measure does not allow its use in the econometric analysis.
5. Note the distinction between disparities in regional growth rates, which are discussed here, and disparities in levels of regional incomes, which, as the literature has shown, kept increasing.
6. Distance from Brussels, defined here relative to the national median distance from Brussels, captures the east–west dichotomy within countries. For most variables, the difference is more accentuated when using the global median (across countries within the same period) as the benchmark. In that case, the shock appears to have been bigger also in more agricultural, larger and less developed economies – suggesting that national economies with such characteristics have performed much worse during that period.
7. For related approaches to the modelling of TFP growth, see, inter alia, Loko and Diouf (Citation2009) and Danquah, Moral-Benito, and Ouattara (Citation2014).
8. In the absence of data on capital stock, the standard approach in the literature (Barro, Citation2001; Agiomirgianakis, Asteriou, & Monastiriotis, Citation2002; Gennaioli, La Porta, De Silanes, & Shleifer, Citation2014) is to proxy capital growth by the investment-to-GDP ratio. The two variables are econometrically indistinguishable when the capital-to-output ratio (and depreciation rate) is constant across the sample (Pritchett, Citation1996). In this analysis differences in capital-to-output ratios are partly controlled for through the use of country-specific time trends. The remaining within-country cross-regional variance is likely to bias the ‘capital growth’ estimates downwards (because they will tend to overestimate capital growth in the more capital-abundant regions), but is unlikely to influence the estimates on the other variables.
9. The F-tests for the difference in the derived coefficients between pairs of periods are as follows (p-values in parentheses): F[c21 = c22] = 202.92 (0.000); F[c21 = c23] = 12.86 (0.0003); F[c21 = c24] = 15.20 (0.0001); F[c22 = c23] = 83.35 (0.000); F[c22 = c24] = 33.16 (0.000); and F[c23 = c24] = 1.57 (0.2096).
10. For a visual depiction of this, see Figure A1 in the supplemental data online, where the four EU association effects are expressed as marginal effects and not as deviations from the ‘grand mean’.
11. Population density, the services share and distance from Brussels now become statistically significant, but the distance from the capital variable loses its significance.
12. As shown in Table A2 in the supplemental data online, in alternative specifications the estimates for the early transition and interim periods have the right sign in all models and are always significant. The results for the pre- and post-accession estimates show more variability: in models that include year-specific fixed effects the pre-accession effect becomes indistinguishable from the sample mean (but still showing a deterioration compared with the interim agreements period) and, partly as a result, the post-accession effect becomes now statistically significant; while in the fixed-effects model (‘within’ estimator) both effects turn positive and statistically significant.
13. As a corollary, Table A3 in the supplemental data online presents a similar analysis using the categorical variables associated with regional characteristics – thus obtaining region-type specific estimates of the impact of these shifts on regional growth. The interaction coefficients in these models are almost always in line with those obtained using the continuous measures, but almost invariably they are estimated less precisely (larger standard errors). This seems to suggest the absence of significant non-linear or threshold effects, at least at the region around the medians of the corresponding distributions.
14. Today, most ENP countries have active trade/cooperation agreements with the EU (Partnership and Cooperation Agreements) and some already move to deeper forms of association (Deep and Comprehensive Free Trade Agreements) which could be paralleled to the Europe agreements that initialled the pre-accession phase in the CEE case.