Using Time-Series Data to Study Civil War Influence in a (Post-) Conflict Economy, Uganda, 1997 - 2006

It has become increasingly clear that pooled samples of time-invariant or time-averaged conflict variables used in cross-country regressions provide a misleading characterization of civil wars in individual countries. Cross-national generalisation is also hampered by the fact that the experience of each country tends to be unique; some suffer total war destruction, others only a slight disturbance. This thesis uses time-series data for the frequency and intensity of civil war violence in econometric models to explain the dynamics of Uganda's (post-) conflict political economy. Three empirical studies are performed using newly constructed monthly indicators of civil war violence reported in local daily newspapers during the period 1997:7 – 2006:6. The studies adopt a general-to-specific modelling approach using an automated model reduction and variable selection algorithm. The main findings are as follows: First, contrary to the widely held view, rebels' offensives are less severe than the government's army's 'peace-keeping' operations, which continually add to the death toll, although not so many people are killed by either protagonist as Uganda's economic performance improves. Second, it is not easy to disentangle the way in which the intensity of the protagonists' joint force affects Uganda's industry: Civil war violence does not seem to be so relevant to the investigation of non-agricultural industry output fluctuations, but it has retardative effects on industries that process agricultural raw materials. At the aggregate level, the results suggest that Uganda's industry generally benefits when the government army appear to be winning the war. Third, the magnitudes of the frequency of protagonists' simultaneous attacks are equal though of opposite effects on the exchange rate market; the Uganda shilling depreciates against the US dollar when rebels attack, but it appreciates when the government's army retaliates. In such a cycle of violence, the government's army's military operation aimed at ending the rebel insurgency has implications for both military and economic policies. But the policy could go beyond military or judicial solutions to the conflict; these do not seem to be working. The finding that the rebels kill fewer people in response to newspaper reports about peace initiatives suggests that the conflict could be resolved non-violently. But the finding that a peaceful resolution to the conflict would retard the long-run growth of industry seems illogical, as it is not in line with prior expectations