Nevertheless the robustness of the coefficients with

This preview shows page 20 - 22 out of 35 pages.

Nevertheless, the robustness of the coefficients with regard to the exclusion of high values enhances the reliance on the validity of the estimates. In both cases, the magnitude of the respective coefficients increases, while the statistical significance decreases if the outliers are excluded. Without the observation of the year 1982 for the region “Ile-de-France”, the estimated coefficient for France is -0.005 per 10 incidents (compared to -0.028 in table 1); the t-value decreases from -4.03 to -3.55 and hence the probability that the true coefficient is zero increases from 0.0 percent to 0.1 percent. The picture is somewhat different in the case of the estimations for the Northern Ireland conflict: though the magnitude of the estimated coefficient is robust, the t-value decreases from -4.00 to -1.11, and the probability that the true coefficient is zero increases from 0.0 percent to 27.2 percent. This latter result indicates that in fact the aforementioned turning point in violence helps to identify the negative effect of the Northern Ireland conflict on individuals’ well-being. (ii) One might expect the negative correlation between terrorism and life satisfaction to be spurious because the state of the economy may be (negatively) correlated with the intensity of terrorist campaigns. If a high level of terrorist activity is generally accompanied by a slackening of the economy, people may report lower life satisfaction because of fearing job reductions rather than terrorist attacks. Such a negative relationship between economic and terroristic activity is probable and can be due to two different causal processes. On the one hand, bad economic conditions and high levels of unemployment may facilitate the recruitment of rank and file activists. On the other hand, terrorism may dampen economic activity. McKittrick and McVea (2001, p. 28) presume the first causal relationship for Northern Ireland when writing that “[i]t is scarcely a coincidence that some of the areas which featured prominently in the troubles […] were among those where […] high unemployment persisted.” However, systematic evidence for the Middle East does not indicate a causal connection from poverty to terrorism (Krueger and Maleckova 2003). The other causal connection is better documented. Over the last few years, economic scholars have analyzed the effects terrorist acts have on various parts of the economy. Empirical research has shown that terrorist acts significantly reduce the number of tourists (see Enders and Sandler 1991 for Spain; Enders et al. 1992 for Greece, Italy and Austria; and Drakos and Kutan 2003 for Greece, Turkey and Israel), lower the inflow of foreign direct investment (see Enders and Sandler 1996 for Spain and Greece), decrease bilateral trade (Nitsch and Schumacher 2002),
20 have significant negative impact on aggregate consumption and savings (see Fielding 2003 for Israel) as well as gross domestic income and stock prices (see Abadie and Gardeazabal 2003 for the Basque Country). In order to control for the simultaneity between high levels of

  • Left Quote Icon

    Student Picture

  • Left Quote Icon

    Student Picture

  • Left Quote Icon

    Student Picture