Growth recovery after civil conflict: a fractional integration approach

Abstract

Using recent econometric techniques based on fractional integration, we find that developing countries recover their economic growth faster than developed countries in response to a shock. Following this methodology, we find that longer civil conflicts are associated with a faster recovery process. We further investigate this issue by exploring correlations with components of GDP, military spending, institutions and aid and find heterogeneous effects of these channels by duration of conflict. Higher government spending is correlated with faster recoveries post longer conflicts, and higher consumption spending is linked to faster recoveries following shorter conflicts. Military spending appears to be driving the government expenditure that makes countries recover from longer conflicts. More democratic institutions are associated with faster recoveries post short wars but slower recoveries following long wars.

Keywords:

• Economic growth recovery
• Civil wars
• Time series

JEL Codes:

• H56
• F43
• O40

Acknowledgements

Comments from the Editor and an anonymous referee are gratefully acknowledged. The usual disclaimer applies.

Disclosure statement

No potential conflict of interest was reported by the authors.

Notes

¹ These authors, using panel data methods found that for a large number of countries economic contractions are not followed by offsetting fast recoveries. They used, however, methods based on integer degrees of differentiation.

² Abadie and Gardeazabal (2003) found that terrorist violence in the Basque Country significantly reduced economic growth relative to a matched control region. Christodoulaki (Forthcoming) examined the cost of the conflict that took place in Greece during 1946–1949. In Rwanda, 20% of the population moved into poverty following the genocide (Justino and Verwimp 2006). Civil wars may also have negative growth spillovers on neighboring countries (Murdoch and Sandler 2004).

³ See Abreu and de Groot (2005) and Hauk and Wacziarg (2009) for a review of the research done on post-war economic growth.

⁴ See Solarin and Sahu (2015) for another paper on the effect of military expenditure on stock market development. Also, see Dimitraki and Ali (2015) for a long run relationship between military expenditure and economic growth in China.

⁵ The issue of why some civil wars last much longer than others is also relevant in the present context. According to Fearon (2004): ‘Civil wars emerging from coups or revolutions tend to be short. Civil wars in Eastern Europe and the former Soviet Union have also tended to be relatively brief, as have anti-colonial wars. By contrast, “sons of the soil” wars that typically involve land conflict between a peripheral ethnic minority and state-supported migrants of a dominant ethnic group are on average quite long-lived. So are conflicts in which a rebel group derives major funding from contraband such as opium, diamonds, or coca’. In a related paper, Collier, Hoeffler, and Söderbom (2004) used hazard functions to determine the duration of civil wars and found that the existence of a few large ethnic groups or forest cover influence the duration of civil wars.

⁶ I(1) implies the presence of a unit root.

⁷ An I(0) process is defined as a covariance stationary process with spectral density function that is positive and finite. It includes the white noise case but also ‘weak’ autocorrelation of the ARMA class.

⁸ See Gil-Alana and Hualde (2009) for a recent review of applications of fractional integration in time series data.

⁹ Nonlinear deterministic trends of the Chebyshev form (Bierens 1997; Cuestas and Gil-Alana 2012) were found to be statistically insignificant in all cases.

¹⁰ We have separated in this figure the countries that present statistically significant estimates of short memory in the growth rates or d > 1 in the logged GDP per capita series). In Appendix 2, Figure A1 displays plots of the GDPPC series for the countries with no trend and war episodes (Afghanistan, Indonesia, Iraq, Liberia, and Sierra Leone), while Figure A2 refers to the countries with no trend and no war episodes; the latter group includes these five countries Bahamas, Haiti, Malta, Trinidad & Tobago, and Zimbabwe. These countries seem to be the exceptions in terms of the mechanism that drives their economic growth.

¹¹ It may be argued that some outliers may be present in the results presented across Figures 1–3. In the cases of Figures 1 and 2, the presence of outliers did not alter the results presented. However, removing countries such as China and Guinea Bissau (in the left-hand-side plots) or China and Iraq (right-hand-side plots), the coefficients became statistically insignificant in all cases.

¹² The average length of civil war duration is approximately 7 years (Collier, Hoeffler, and Söderbom 2004).

¹³ Available at www.aiddata.org.