69 Conditions That Affect the Impact of Trade Shocks on Conflict Table 2.1 Under What Conditions Are the Marginal Effects of Trade Shocks Not Significant? Type Range Px Point source Grievance Economic inequality gini_net gini_market Ethnic polarization Ethnic fractionalization Religious fractionalization Religious polarization Any conflict in last 10 yrs Contin. Contin. Contin. Contin. Contin. Contin. Contin. Dummy [22;65] [15;75] [17;80] [0;1] [0;1] [0;0.7] [0;0.96] [0;1] <26 or >58 <32 <37 <0.37 <0.28 No diff. No diff. 0 <46 or >52 <48 <52 <0.42 <0.30 <0.28 <0.49 0 Political institutions Elections (t-1) Federal Government Polity2 Dummy Dummy Contin. [0;1] [0;1] [−10;10] No diff. 1 No diff. Elections Federal >5 Parliamentary democracy Programmatic party Age of party in office Dummy 3 groups 3 groups [0;1] [0;3] [1;191] No diff. < or > 2nd tercile No diff. No diff. No diff. No diff. Governance Law Military Accountability Corruption Bureaucratic Composite index Contin. Contin. Contin. Contin. Contin. Contin. [0;6] [0;6] [0;6] [0;6] [0;4] [12;84.7] No diff. No diff. >3 >2.5 >1.5 No diff. No diff. No diff. >2.2 No diff. No diff. No diff. Neighbors Neighbors’ conflict (any) Share trade neighbors RTA Share trade neighbors Dummy Contin. Contin. [0;1] [0;0.8] [0;0.9] 0 >0.35 No diff. 0 >0.08 >0.24 Price transmission NRA (output + input) 3 groups [−0.9;1.4] < or > 2nd tercile No diff. NRA (output) 3 groups [−0.9;3.5] < or > 2nd tercile No diff. Note: The column range indicates the range of values taken by each variable in our sample; the column type groups the variables into three types, i.e. continuous, dummy (0 or 1), and 3 groups (the variable is split into three mutually exclusive continuous groups of values). The latter category is included for those variables for which the marginal effect of Px on conflict appeared to be non linear. The other columns report the values of the interaction for which the marginal effect of the relevant trade variable becomes not significantly different from zero (at the 10 percent level); “No diff.” indicates either that the marginal effect is always or never statistically different from zero across the distribution of the interactions’ values. The variables are defined in table A.3. from zero once the accountability index rises above 3. Similarly, the upward sloping curve in the lower panel indicates that the effect of the price index on conflict becomes larger for observations characterized by at least one conflict in contiguous countries. In fact, the effect of the price index (while positive) is not significantly different from zero for countries in years that they have peaceful neighbors. These results can also be used to determine the relative importance of the conditioning factors in explaining the resilience to trade-related changes. Trading Away from Conflict • http://dx.doi.org/10.1596/978-1-4648-0308-6