The Economics of Secession: Analysing the Economic Impact of the Collapse of the Former Yugoslavia

Abstract

This paper looks at the economic impact of secession through the lens of the disintegration of the former Yugoslavia. It uses an econometric analysis covering the period between 1956 and 2011—including a series of factors linked to the independence process, socio-economic and structural controls, and the level of development—in order to assess whether (a) breaking away from the former Yugoslavia delivered an ‘independence dividend’ to the newly independent countries and whether (b) independence had a more favourable impact in richer, rather than poorer territories. The results of the analysis underline that there has been no favourable economic impact of secession and that how secession was achieved is key in understanding the subsequent economic performance of the newly independent countries. In cases of secession without conflict, independence did not have a noticeable impact on ensuing economic performance. Secession achieved by conflict, by contrast, seriously dented growth prospects.

Extracto

En este artículo se analiza el impacto económico de la secesión a través del caso de la desintegración de la antigua Yugoslavia. Mediante un análisis econométrico para el periodo 1956–2011—que comprende una serie de factores ligados al proceso de independencia, así como controles socioeconómicos, estructurales y del nivel de desarrollo—se estudia (a) hasta qué punto la separación de un país de la antigua Yugoslavia trajo consigo un ‘dividendo ligado a la independencia’ y (b) si el proceso de independencia tuvo consecuencias económicas más favorables para los Estados más ricos de la antigua Yugoslavia. Los resultados ponen de manifiesto que no ha existido un impacto económico favorable ligado a la independencia y que la forma en que se desarrolló el proceso de secesión es clave para entender el desempeño económico de los países que emergieron de la antigua Yugoslavia. En aquellos casos en que la secesión tuvo lugar sin conflicto, la independencia no tuvo un efecto notable sobre la trayectoria económica posterior. La secesión marcada por conflicto, en cambio, limitó de manera clara al crecimiento económico posterior.

摘要

本文透过前南斯拉夫解体之视角，检视国家分裂的经济影响。本文运用涵盖自1956年至2011年期间的的计量经济分析——包含与独立过程、社会经济和结构控制，以及发展程度相关的一系列因素，评估a）从前南斯拉夫分裂出来，是否为新兴的独立国家带来了‘独立红利’，以及b）在较富裕的领域，独立是否产生较有利的影响，而对较贫困的领域则较不利。分析结果强调，分裂并没有有利的经济影响，而达到分裂的的方式，则是理解新兴独立国家随后的经济表现的关键。在没有冲突的国家分裂案例之中，独立对于日后的经济表现，并未产生显着的冲击。反之，透过冲突达到的国家分裂，则严重地削弱成长的前景

Résumé

Cet article cherche à examiner l'impact économique de la sécession sous l'optique du démantèlement de l'ex-Yougoslavie. On emploie une analyse économétrique pour la période allant de 1956 à 2011—y compris une série de facteurs liés au processus d'indépendance, aux contrôles socioéconomiques et structurels, et au niveau de développement—pour évaluer (a) si leur séparation de l'ex-Yougoslavie a fourni ‘un dividende d'autonomie’ aux nouveaux États indépendants et (b) si l'indépendance a touché plus favorablement les territoires plus riches que les territoires plus pauvres. Les résultats de l'analyse soulignent que la sécession n'a eu aucun impact économique favorable et que la manière dont la sécession a été réalisée est indispensable à la compréhension de la performance économique ultérieure des nouveaux États indépendants. Dans les cas où la sécession était réalisée sans conflit, il s'avère que l'indépendance n'avait pas eu d'impact significatif sur la performance économique qui s'en est suivie. Par contraste, la sécession réalisée suite à un conflit a été sérieusement préjudiciable aux perspectives de croissance.

Keywords:

• secession
• independence
• economic growth
• conflict
• Yugoslavia
• Europe

Notes

1. From hereon just ‘Bosnia’.

2. For reference, the same ratio was close to 14:1 in 2009 (EBRD, 2010).

3. As in most wars, the number of causalities in the different wars of the former Yugoslavia is shrouded in controversy. Different sources—sometimes because of a political bias, others just as a result of the sheer difficulty of accurately measuring War Deaths—provide very different numbers of casualties. Hence, in order to make sure that the results of including War Deaths_it are not affected by the war intensity variable chosen, we have created a second variable of war intensity. This alternative variable includes data stemming from a large number of written and online sources, led by Tabeau's (2009) report on ‘Casualties of the 1990s wars in the former Yugoslavia (1991–1999)’. The differences between this variable and that sourced from the PRIO (2009) Battle Deaths Database variable is non-trivial: whereas the PRIO (2009) Battle Deaths Database records a total of 63,383 War Deaths in the different wars, our alternative measure more than doubles that number, raising the toll to 135,031. Both estimates are, in any case, considerably lower than the majority of those reported in the former Yugoslavia and taking just the Bosnian war into account (see Tabeau and Bijak 2005 for a detailed analysis of this issue). However, the inclusion of the alternative war intensity variable in the model leaves the coefficients of the estimations virtually unchanged. These results can be provided upon request.

4. Sanctions were imposed by the UN Security Council on Serbia and Montenegro in 1992–1996 (UN 1996) and again (but excluding Kosovo) by the EU and USA in 1999–2000. Other sanctions were seen as less important and discounted by the dummy. Although Serb controlled parts of Croatia and Bosnia were also technically sanctioned in 1992–1996, this covered only part of their territories and the dummy was not applied.

5. TRADE_t = (M_t + E_t)/GDP_t, where M_t is the value of a country's imports at a given time t; E_t is the equivalent value of its exports; and GDP_t is denoted in the same currency. Note that the scores are calculated for sovereign countries—that is, not for individual republics—for two reasons: firstly, since it is not clear what would be meant by saying that one republic (as a subnational entity) is more ‘open’ than another, since foreign trade (by definition) occurs between countries; and secondly, since reliable disaggregated data on regional imports and exports were not available.

6. Scores are interpreted as the probability that two individuals randomly picked from a population belong to different ethnic groups (Fearon 2003). Thus, low scores indicate ethnic homogeneity; high scores imply fractionalisation.

7. A detailed discussion of the problems and reliability of different estimates in this field is provided by Brunborg (2002). Wherever possible, the present study uses UNHCR data.