Why do radical revisionist domestic ideologies result in revisionist international demands? Why are dictatorships led by extremists more prone to targeting democracies in interstate conflict? While the pattern of extremism and conflict is not much questioned, the causal mechanism is not thoroughly explained. A common assumption is that in nondemocracies ideological leaders pursue revisionist goals because of personal normative preferences and that they are able to use conflict because domestically the agenda is entirely theirs to set. This conflicts with the assumption that the primary motive of leaders is to remain in power. The costly pursuit of extraterritorial ambitions detracts from the material resources available to payoff support coalitions and otherwise keep the dictator in power. So, why do nondemocratic leaders pursue costly ideological goals abroad rather than efficiently use those resources for payoffs at home? I argue that ideological dictators use revisionist demands to send signals of ideological commitment to their most important domestic supporters. This article demonstrates through an analysis of dyadic MIDS and a monadic analysis of the first use of violent force that it is rational signaling behavior, and not the personal ideological preferences of leaders, which links domestic revisionist ideology to international revisionist conflict.
Keywords
autocracy promotion, conflict, ideology, revisionist states
Hypothesis H1 (The Rational Autocrat): Support Groups Matter
Theoretical basis
W or winning coalition size is proposed by Selectorate Theory as the major driver of state behavior in both democracies and nondemocracies. (Bueno de Mesquita and Smith, multiple dates)
Institutional factors in the arrangement of government have been proposed as an important factor. Institutionalist explanations are primarily concerned with constraints on the leader, often characterized as constrained versus personalist leadership. Secondary consideration is given to the effect of domestic audiences including military versus civilian, and single party versus other types. (Geddes at al., Weeks, Boix and Svolik, others)
Military support matters because accepting military support creates a moral hazard for dictators, yielding large areas of policy control to the military. (Svolik)
Ideology matters in domestic policy because accepting ideological support as a candidate for dictator can alienate the support of other elites, leaving the dictator captured by ideological interests. (Note the similarity to Svolik’s argument regarding the military.) The result in domestic contexts is, with some ideologies, the use of expropriation by leaders to signal ideological commitment to these groups. For the leader, this has utility only while there are reasonable assets to expropriate and the effectiveness of acts of expropriation declines over time. (Albertus and Menaldo 2012, Albertus 2015)
Albertus and Menaldo offer a solid case for the capture of regimes by ideological support groups in the domestic context. As there are less properties to expropriate and as the time since expropriation recedes, it becomes necessary for the ideologically captured leader to find other means of signaling commitment to these supporters. The option presented by selectorate theory, material payoffs to the minimal winning coalition, may not be the most feasible for two reasons. First, the larger coalition of influentials, the ideological support base, likely wields sufficient power to remove a minimal winning coalition if dissatisfied enough and replace it, and paying off this larger group would be prohibitively expensive. Second, the ideological supporters may be more concerned with ideological commitment than material payoffs. One alternative, which may even be less expensive than material payoffs, is turning the revisionist ideology abroad through revisionist MIDS.
Does the explicit ideology or implicit ideological association of the support groups matter for international behavior specifically the initiation of revisionist MIDS and the targeting of democracies?
H1a: Autocratic states with higher levels of regime support from groups associated with revisionist ideologies will be more likely to originate a revisionist MID than other autocratic states all else equal.
H1b: Autocratic states with higher levels of regime support associated with revisionist ideologies will be more likely to originate MIDS targeting democracies than other autocratic states, all else being equal.
H1c: Autocratic states with higher levels of domestic ideological support will be more likely to engage in the First Use of Violent Force than other autocratic states, all else equal.
Hypothesis H1d: Autocratic states with higher levels of domestic ideological support will be more likely to engage in the First Use of Violent Force in MIDS against a democratic target than other autocratic states, all else equal.
Hypothesis H2 (The Ideological Autocrat): Regime leadership ideology matters
Basis of the argument
That the ideological character of a regime is associated with increased conflict levels is widely accepted with a number of proposed mechanisms including Messianic Autocrats, coincidental association with W or other institutional characteristics, preventive war, and specifically ideological conflicts.
These common assumptions are not well tested.
This section begins to fill this gap with a well controlled empirical test of the influence of leadership ideology.
H2a: Autocratic states with higher levels of revisionist domestic leadership will will be more likely to originate a revisionist MID than other autocratic states, all else equal.
Show the code
h2a_full <-glm(sidea_revisionist ~ sidea_revisionist_domestic + sidea_national_military_capabilities + sideb_national_military_capabilities + sidea_e_miinteco + cold_war + sidea_revisionist_domestic + sidea_national_military_capabilities + sideb_national_military_capabilities + sidea_military_support + sidea_winning_coalition_size + cold_war + t + t2 + t3 + sidea_muslim, family =binomial(link ="logit"), data = autocracies)h2a_multi_ideology <-glm(sidea_revisionist ~ sidea_nationalist_revisionist_domestic +sidea_socialist_revisionist_domestic +sidea_reactionary_revisionist_domestic +sidea_separatist_revisionist_domestic +sidea_religious_revisionist_domestic + sidea_national_military_capabilities + sideb_national_military_capabilities + sidea_military_support + sidea_winning_coalition_size + cold_war + t + t2 + t3 + sidea_muslim, family =binomial(link ="logit"), data = autocracies) stargazer(h2a_full,h2a_multi_ideology, type ="text")
H2b: Autocratic states with higher levels of revisionist domestic leadership will will be more likely to originate a revisionist MID targeting a democracy than other autocratic states, all else equal.
Show the code
h2b_full <-glm(sidea_targets_democracy ~ sidea_revisionist_domestic + sidea_national_military_capabilities + sideb_national_military_capabilities + sidea_e_miinteco + cold_war + sidea_revisionist_domestic + sidea_national_military_capabilities + sideb_national_military_capabilities + sidea_military_support + sidea_winning_coalition_size + cold_war + t + t2 + t3 + sidea_muslim, family =binomial(link ="logit"), data = autocracies) h2b_multi_ideology <-glm(sidea_targets_democracy ~ sidea_nationalist_revisionist_domestic +sidea_socialist_revisionist_domestic +sidea_reactionary_revisionist_domestic +sidea_separatist_revisionist_domestic +sidea_religious_revisionist_domestic + sidea_national_military_capabilities + sideb_national_military_capabilities + sidea_military_support + sidea_winning_coalition_size + cold_war + t + t2 + t3 + sidea_muslim, family =binomial(link ="logit"), data = autocracies) stargazer(h2b_full,h2b_multi_ideology, type ="text")
H2c: Autocratic states with higher levels of revisionist domestic leadership will will be more likely to be the first to use violent force when involved in a MID than other autocratic states, all else equal.
Show the code
h2c_multi_ideology <-glm(first_use_of_force ~ nationalist_revisionist_domestic + socialist_revisionist_domestic +reactionary_revisionist_domestic +separatist_revisionist_domestic +religious_revisionist_domestic + national_military_capabilities + military_support + winning_coalition_size + cold_war + t + t2 + t3 + muslim, family =binomial(link ="logit"), data = autocracies_m)stargazer(h2c_multi_ideology, type ="text")
H2d: Autocratic states with higher levels of revisionist domestic leadership will will be more likely to be the first to use violent force when involved in a MID targeting a democracy than other autocratic states, all else equal.
Show the code
h2d_multi_ideology <-glm(first_use_of_force_targeting_democracy ~ nationalist_revisionist_domestic + socialist_revisionist_domestic +reactionary_revisionist_domestic +separatist_revisionist_domestic +religious_revisionist_domestic + national_military_capabilities + military_support + winning_coalition_size + cold_war + t + t2 + t3 + muslim, family =binomial(link ="logit"), data = autocracies_m)stargazer(h2d_multi_ideology, type ="text")
Hypothesis 3 (The Rational Autocrat): Leader ideology is mediated by leader’s dependence on ideology associated support groups
Theoretical basis
It is accepted as noted above that non-ideological support groups matter to the behavior of dictators.
A variety of explanations have been offered.
Albertus and Menado demonstrate the capture of leaders by socialist ideological groups in domestic politics.
The new theoretical model offered and tested here is that the behavior of ideological leaders can be explained by the need to secure the continued loyalty of ideological support groups in the domestic audience.
Simple mediation 1 - Religious support groups associated with religious ideology
lavaan 0.6-19 ended normally after 133 iterations
Estimator DWLS
Optimization method NLMINB
Number of model parameters 18
Used Total
Number of observations 316435 651456
Model Test User Model:
Standard Scaled
Test Statistic 0.00000 0.00000
Degrees of freedom 0 0
Parameter Estimates:
Parameterization Delta
Standard errors Robust.sem
Information Expected
Information saturated (h1) model Unstructured
Regressions:
Estimate Std.Err z-value P(>|z|)
sidea_targets_democracy ~
sd_rlgs_sp (b) 5.16652 0.01862 277.40548 0.00000
sd_rlgs_r_ (c) 0.81505 0.06690 12.18299 0.00000
sd_ntnl_m_ 9.19118 0.75150 12.23038 0.00000
sdb_ntnl__ 7.37365 0.35062 21.03042 0.00000
sd_mltry_s 0.04548 0.10208 0.44551 0.65595
sd_wnnng__ -0.12410 0.22785 -0.54465 0.58599
sidea_mslm -0.12782 0.04819 -2.65262 0.00799
t -0.00837 0.00178 -4.68872 0.00000
sidea_religious_support ~
sd_rlgs_r_ (a) 0.01136 0.00139 8.14497 0.00000
sd_ntnl_m_ -0.58238 0.09196 -6.33333 0.00000
sdb_ntnl__ -0.00253 0.01606 -0.15770 0.87470
sd_mltry_s 0.07352 0.00152 48.50601 0.00000
sd_wnnng__ 0.14691 0.00318 46.22300 0.00000
sidea_mslm 0.08901 0.00089 99.52746 0.00000
t 0.00100 0.00004 26.18385 0.00000
Intercepts:
Estimate Std.Err z-value P(>|z|)
.sid_rlgs_spprt -0.02259 0.00255 -8.86723 0.00000
Thresholds:
Estimate Std.Err z-value P(>|z|)
sd_trgts_dmc|1 4.11378 0.14769 27.85349 0.00000
Variances:
Estimate Std.Err z-value P(>|z|)
.sd_trgts_dmcrc 0.11974
.sid_rlgs_spprt 0.03298 0.00010 322.04502 0.00000
Defined Parameters:
Estimate Std.Err z-value P(>|z|)
indirect 0.05868 0.00723 8.12043 0.00000
total 0.87373 0.06635 13.16920 0.00000
Simple mediation 2 - Rural worker support associated with religious ideology
lavaan 0.6-19 ended normally after 123 iterations
Estimator DWLS
Optimization method NLMINB
Number of model parameters 18
Used Total
Number of observations 316435 651456
Model Test User Model:
Standard Scaled
Test Statistic 0.00000 0.00000
Degrees of freedom 0 0
Parameter Estimates:
Parameterization Delta
Standard errors Robust.sem
Information Expected
Information saturated (h1) model Unstructured
Regressions:
Estimate Std.Err z-value P(>|z|)
sidea_targets_democracy ~
sd_rrl_wr_ (b) 4.24038 0.01173 361.47524 0.00000
sd_rlgs_r_ (c) 0.22380 0.06632 3.37472 0.00074
sd_ntnl_m_ 5.13527 0.62740 8.18498 0.00000
sdb_ntnl__ 7.28857 0.34961 20.84778 0.00000
sd_mltry_s 0.64557 0.10171 6.34726 0.00000
sd_wnnng__ 0.71352 0.22648 3.15046 0.00163
sidea_mslm 0.83932 0.04780 17.56080 0.00000
t -0.00577 0.00179 -3.22325 0.00127
sidea_rural_worker_support ~
sd_rlgs_r_ (a) 0.15327 0.00111 137.47332 0.00000
sd_ntnl_m_ 0.24692 0.04238 5.82564 0.00000
sdb_ntnl__ 0.01698 0.01509 1.12493 0.26062
sd_mltry_s -0.05194 0.00133 -39.08585 0.00000
sd_wnnng__ -0.01854 0.00269 -6.89514 0.00000
sidea_mslm -0.11962 0.00102 -117.05936 0.00000
t 0.00060 0.00004 14.60693 0.00000
Intercepts:
Estimate Std.Err z-value P(>|z|)
.sd_rrl_wrkr_sp 0.18757 0.00238 78.68549 0.00000
Thresholds:
Estimate Std.Err z-value P(>|z|)
sd_trgts_dmc|1 5.02587 0.14719 34.14498 0.00000
Variances:
Estimate Std.Err z-value P(>|z|)
.sd_trgts_dmcrc 0.38914
.sd_rrl_wrkr_sp 0.03397 0.00012 278.24323 0.00000
Defined Parameters:
Estimate Std.Err z-value P(>|z|)
indirect 0.64993 0.00472 137.58710 0.00000
total 0.87373 0.06635 13.16920 0.00000
Simple mediation 3 - Party elite support associated with socialist ideology
Show the code
simpleMediation3 <-'sidea_targets_democracy ~ b * sidea_party_elite_support + c * sidea_socialist_revisionist_domestic + sidea_national_military_capabilities + sideb_national_military_capabilities + sidea_military_support + sidea_winning_coalition_size + sidea_muslim + t sidea_party_elite_support ~ a * sidea_socialist_revisionist_domestic + sidea_national_military_capabilities + sideb_national_military_capabilities + sidea_military_support + sidea_winning_coalition_size + sidea_muslim + t indirect := a * b total := c + (a * b) 'fit_s3 <-sem(model = simpleMediation3, ordered ="sidea_targets_democracy", data = autocracies)
Warning: lavaan->lav_samplestats_step2():
correlation between variables sidea_party_elite_support and
sidea_targets_democracy is (nearly) 1.0
Show the code
summary(fit_s3, nd =5)
lavaan 0.6-19 ended normally after 124 iterations
Estimator DWLS
Optimization method NLMINB
Number of model parameters 18
Used Total
Number of observations 316435 651456
Model Test User Model:
Standard Scaled
Test Statistic 0.00000 0.00000
Degrees of freedom 0 0
Parameter Estimates:
Parameterization Delta
Standard errors Robust.sem
Information Expected
Information saturated (h1) model Unstructured
Regressions:
Estimate Std.Err z-value P(>|z|)
sidea_targets_democracy ~
sd_prty_l_ (b) 4.07983 0.00691 590.27036 0.00000
sd_sclst__ (c) -0.50368 0.05412 -9.30665 0.00000
sd_ntnl_m_ -3.02908 0.61134 -4.95482 0.00000
sdb_ntnl__ 7.57250 0.35163 21.53520 0.00000
sd_mltry_s 0.83299 0.10295 8.09086 0.00000
sd_wnnng__ -1.04555 0.23483 -4.45233 0.00001
sidea_mslm 1.41647 0.05109 27.72746 0.00000
t 0.01597 0.00176 9.07249 0.00000
sidea_party_elite_support ~
sd_sclst__ (a) 0.31185 0.00125 248.61373 0.00000
sd_ntnl_m_ 2.25501 0.04594 49.08583 0.00000
sdb_ntnl__ -0.02576 0.02154 -1.19600 0.23170
sd_mltry_s -0.08508 0.00152 -56.09749 0.00000
sd_wnnng__ 0.51043 0.00354 144.29012 0.00000
sidea_mslm -0.21496 0.00119 -180.19269 0.00000
t -0.00484 0.00006 -77.05461 0.00000
Intercepts:
Estimate Std.Err z-value P(>|z|)
.sd_prty_lt_spp 0.50524 0.00328 154.13512 0.00000
Thresholds:
Estimate Std.Err z-value P(>|z|)
sd_trgts_dmc|1 6.59694 0.16092 40.99500 0.00000
Variances:
Estimate Std.Err z-value P(>|z|)
.sd_trgts_dmcrc 0.01680
.sd_prty_lt_spp 0.05907 0.00018 319.53101 0.00000
Defined Parameters:
Estimate Std.Err z-value P(>|z|)
indirect 1.27231 0.00525 242.25737 0.00000
total 0.76863 0.05383 14.27940 0.00000
Simple mediation 4 - Military support associated with nationalist ideology
Regression Model and Mediation Model
Because military support is necessary for every ideology, the basic correlation between nationalism and military support may be spurious. With the inclusion of the full controls used in the previous models, the relationship between nationalist ideology and military reverses sign and the indirect effect is therefore none.
Show the code
mediator_nationalism <-lm(sidea_military_support ~ sidea_nationalist_revisionist_domestic + sidea_national_military_capabilities + sideb_national_military_capabilities + sidea_winning_coalition_size + sidea_muslim + t, data = autocracies)stargazer(mediator_nationalism, type ="text")
simpleMediation4 <-'sidea_targets_democracy ~ b * sidea_military_support + c * sidea_nationalist_revisionist_domestic +sidea_national_military_capabilities + sideb_national_military_capabilities + sidea_winning_coalition_size + sidea_muslim + t sidea_military_support ~ a * sidea_nationalist_revisionist_domestic + sidea_national_military_capabilities + sideb_national_military_capabilities + sidea_winning_coalition_size + sidea_muslim + t indirect := a * b total := c + (a * b) 'fit_s4 <-sem(model = simpleMediation4, ordered ="sidea_targets_democracy", data = autocracies)summary(fit_s4, nd =5)
lavaan 0.6-19 ended normally after 115 iterations
Estimator DWLS
Optimization method NLMINB
Number of model parameters 16
Used Total
Number of observations 316435 651456
Model Test User Model:
Standard Scaled
Test Statistic 0.00000 0.00000
Degrees of freedom 0 0
Parameter Estimates:
Parameterization Delta
Standard errors Robust.sem
Information Expected
Information saturated (h1) model Unstructured
Regressions:
Estimate Std.Err z-value P(>|z|)
sidea_targets_democracy ~
sd_mltry_s (b) 0.33672 0.06670 5.04858 0.00000
sd_ntnls__ (c) 0.87738 0.06305 13.91642 0.00000
sd_ntnl_m_ 6.42895 0.59149 10.86900 0.00000
sdb_ntnl__ 7.40519 0.34049 21.74857 0.00000
sd_wnnng__ 0.55011 0.20513 2.68181 0.00732
sidea_mslm 0.38241 0.04459 8.57549 0.00000
t -0.00484 0.00189 -2.56185 0.01041
sidea_military_support ~
sd_ntnls__ (a) -0.06088 0.00171 -35.66187 0.00000
sd_ntnl_m_ -0.68682 0.06346 -10.82205 0.00000
sdb_ntnl__ -0.05729 0.02231 -2.56836 0.01022
sd_wnnng__ -0.46616 0.00416 -112.18674 0.00000
sidea_mslm -0.02796 0.00126 -22.12228 0.00000
t -0.00638 0.00008 -79.57805 0.00000
Intercepts:
Estimate Std.Err z-value P(>|z|)
.sd_mltry_spprt 1.05282 0.00374 281.58044 0.00000
Thresholds:
Estimate Std.Err z-value P(>|z|)
sd_trgts_dmc|1 4.30006 0.12688 33.89171 0.00000
Variances:
Estimate Std.Err z-value P(>|z|)
.sd_trgts_dmcrc 0.99175
.sd_mltry_spprt 0.07277 0.00025 286.63663 0.00000
Defined Parameters:
Estimate Std.Err z-value P(>|z|)
indirect -0.02050 0.00410 -4.99417 0.00000
total 0.85688 0.06290 13.62354 0.00000
Alternate Hypothesis - (The Messianic Autocrat): Dynamic leadership matters
A1 - If the alternate hypothesis of the Messianic Autocrat is correct, then dynamic leader qualities will increase the likelihood of conflict behavior, all else equal.
A1a
Show the code
alternate1full <-glm(sidea_revisionist ~ sidea_dynamic_leader + sidea_revisionist_domestic + sidea_national_military_capabilities + sideb_national_military_capabilities + sidea_military_support + sidea_winning_coalition_size + cold_war + t + t2 + t3 + sidea_muslim, family =binomial(link ="logit"), data = autocracies) alternate2full <-glm(sidea_targets_democracy ~ sidea_dynamic_leader + sidea_revisionist_domestic + sidea_national_military_capabilities + sideb_national_military_capabilities + sidea_military_support + sidea_winning_coalition_size + cold_war + t + t2 + t3 + sidea_muslim, family =binomial(link ="logit"), data = autocracies) stargazer(alternate1full,alternate2full, type ="text")
A2 - If the alternate hypothesis of Messianic Autocrats is correct, there will be a positive interaction effect between leadership ideology and dynamic leadership on conflict behavior, all else equal.
A2a
Show the code
alternate1int <-lm(sidea_revisionist ~ sidea_dynamic_leader * sidea_revisionist_domestic + sidea_national_military_capabilities + sideb_national_military_capabilities + sidea_military_support + sidea_winning_coalition_size + cold_war + t + t2 + t3 + sidea_muslim, data = autocracies) alternate2int <-lm(sidea_targets_democracy ~ sidea_dynamic_leader * sidea_revisionist_domestic + sidea_national_military_capabilities + sideb_national_military_capabilities + sidea_military_support + sidea_winning_coalition_size + cold_war + t + t2 + t3 + sidea_muslim, data = autocracies) stargazer(alternate1int,alternate2int, type ="text")
plot_model(h2a_full, type ="pred", terms ="sidea_revisionist_domestic", title ="GLM model Predicted Probabilities")
Show the code
## plot predicted probabilities for h1alternatefull using sjplot plot_modelplot_model(alternate1full, type ="pred", terms =c("sidea_dynamic_leader","sidea_revisionist_domestic"), title ="GLM model Predicted Probabilities")
Show the code
## plot predicted probabitilies for h2alternatefull using sjplot plot_modelplot_model(alternate1int, type ="pred", terms =c("sidea_dynamic_leader","sidea_revisionist_domestic"), title ="Linear model with interaction effects \n Predicted Probabilities")
Show the code
## plot predicted probabilities for alternate2full using sjplot plot_modelplot_model(alternate2full, type ="pred", terms =c("sidea_dynamic_leader","sidea_revisionist_domestic"), title ="GLM model Predicted Probabilities")
Show the code
## plot predicted probabitilies for alternate2int using sjplot plot_modelplot_model(alternate2int, type ="pred", terms =c("sidea_dynamic_leader","sidea_revisionist_domestic"), title ="Linear model with interaction effects \n Predicted Probabilities")
Show the code
plot_model(alternate2int, type ="pred", terms =c("sidea_revisionist_domestic", "sidea_dynamic_leader"), title ="Linear model with interaction effects \n Predicted Probabilities")
A3 - leader ideology is mediated by dynamic leadership
Simple Mediation 3 - Dynamic Leader (Alternative)
lavaan 0.6-19 ended normally after 146 iterations
Estimator DWLS
Optimization method NLMINB
Number of model parameters 18
Used Total
Number of observations 316435 651456
Model Test User Model:
Standard Scaled
Test Statistic 0.00000 0.00000
Degrees of freedom 0 0
Parameter Estimates:
Parameterization Delta
Standard errors Robust.sem
Information Expected
Information saturated (h1) model Unstructured
Regressions:
Estimate Std.Err z-value P(>|z|)
sidea_targets_democracy ~
sd_dynmc_l (b) 0.96328 0.00155 620.57621 0.00000
sd_rlgs_r_ (c) -0.24617 0.06615 -3.72149 0.00020
sd_ntnl_m_ 8.23405 0.64030 12.85972 0.00000
sdb_ntnl__ 7.46466 0.35483 21.03715 0.00000
sd_mltry_s 0.09544 0.10187 0.93687 0.34882
sd_wnnng__ 0.03970 0.22681 0.17503 0.86105
sidea_mslm 0.06366 0.04802 1.32584 0.18489
t -0.00490 0.00181 -2.70783 0.00677
sidea_dynamic_leader ~
sd_rlgs_r_ (a) 1.16258 0.00639 182.06424 0.00000
sd_ntnl_m_ -2.12997 0.18268 -11.65967 0.00000
sdb_ntnl__ -0.10806 0.09101 -1.18732 0.23510
sd_mltry_s 0.34248 0.00772 44.38340 0.00000
sd_wnnng__ 0.61791 0.01569 39.37131 0.00000
sidea_mslm 0.27864 0.00526 52.98877 0.00000
t 0.00174 0.00031 5.59335 0.00000
Intercepts:
Estimate Std.Err z-value P(>|z|)
.side_dynmc_ldr 0.38308 0.01505 25.45145 0.00000
Thresholds:
Estimate Std.Err z-value P(>|z|)
sd_trgts_dmc|1 4.59951 0.14755 31.17245 0.00000
Variances:
Estimate Std.Err z-value P(>|z|)
.sd_trgts_dmcrc 0.03293
.side_dynmc_ldr 1.04220 0.00287 363.70762 0.00000
Defined Parameters:
Estimate Std.Err z-value P(>|z|)
indirect 1.11990 0.00608 184.09289 0.00000
total 0.87373 0.06635 13.16920 0.00000
H4: Multiple mediation models
Hypothesis H4a: Mediation of radical religious leadership ideology by religious support and dynamic leader.
Multiple mediation 1 - Religious support (Main Hypotheses) and dynamic leader (Alternative Hypothesis)
I include minimal winning coalition size, computed using the Bueno de Mesquita and Smith (2022) formula and the current VDem data, as a covariate in all models.
Institutional characteristics
As noted by Bueno de Mesquita and Smith (2022) and their critics Clarke and Stone (2008), W, or minimal winning coalition size, is “reducible to institutional incentives.” The measure of W used by Bueno de Mesquita and Smith, and replicated as a control in this study, is based on 4 institutional variables from the VDem data and has been assessed against a variety of alternative institutional measures. I will not reinvent the wheel here.
Probability of being attacked by a democracy
I run a logistic regression using the same sample and independent variables as H2 with the dependent variable Targeted by Democracy.
With the same sample and independent variables, the confidence intervals of the coefficients are comparable.
The coefficient for Targets Democracy is approximately two times the size of the coefficient for Targeted by Democracy and the confidence intervals do not overlap.
Future work
The results show that there is variation in behavior between revisionist ideology types. They also show variation at the four levels of behavior tested: revisionist MIDS initiation, revisionist MIDS initiation against democracies, first use of violent force in a MID, and first use of violent force in a MID against democracies. Especially at the highest level, but the one with the smallest sample size, the results show that the Rational Autocrat is less likely to initiate a MID against a democracy than the Messianic Autocrat. This is consistent with the idea that the Rational Autocrat is more likely to be deterred by the higher costs of attacking a democracy. At lower levels, the Rational Autocrat is more likely and the Messianic Autocrat less likely to initiate a MID. The theoretical explanation here is that at lower levels of aggression, the benefit of maintaining ideological support is greater than the risk of antagonizing other states including democratic ones. At higher levels, the support groups themselves become less supportive and the greater ability of the Messianic autocrat to influence domestic groups is more important. Future work could further explore these relationships in much greater detail, especially through the use of case studies.
Is there a difference between aggressive behavior toward non-democracies and aggressive behavior toward democracies. A higher level of aggression toward democracies would indicate two important things - that the signal of attacking a democracy is stronger and that there is an active program of conflict as autocracy promotion not merely a predisposition toward conflict generally. The complicating issue is that one hypothetical explanation of democratic peace theory is that democracies as more likely to win wars. If this is the case, a simple test based on proportions would not be falsifiable since there is a ready alternate explanation for a broader tendency to target nondemocracies. A good test will require a rather complicated model to take all known and proposed differences between democratic and nondemocratic targets into account.
Does the ideological character of a regime influence the size of minimal winning coalitions? Does it influence ths size of the intermediate group, the Influentials?
Where do these ideological payoffs fit in the Selectorate Theory of public and private goods? Should they be considered as another type of good entirely. Are they club goods, nonrivalrous but only meaningful to the ideological supporter and so quite exclusive?
Do payoffs to winning coalitions vary between revisionist regimes which engage in revisionist MIDS and other nondemocratic regimes? Lower payoffs to Ws in the former case would strengthen the theory. Considerable new data would have to be developed at a fairly high degree of difficulty given the non-transparent nature of these regimes.
Authors cited
Complete citations are included in the manuscript and will be updated here as this draft is completed.
@unpublished{hanna,
author = {Hanna, Tom},
title = {Theory: {Brief} {Outline,} {Tables,} and {Graphs}},
langid = {en},
abstract = {Why do radical revisionist domestic ideologies result in
revisionist international demands? Why are dictatorships led by
extremists more prone to targeting democracies in interstate
conflict? While the pattern of extremism and conflict is not much
questioned, the causal mechanism is not thoroughly explained. A
common assumption is that in nondemocracies ideological leaders
pursue revisionist goals because of personal normative preferences
and that they are able to use conflict because domestically the
agenda is entirely theirs to set. This conflicts with the assumption
that the primary motive of leaders is to remain in power. The costly
pursuit of extraterritorial ambitions detracts from the material
resources available to payoff support coalitions and otherwise keep
the dictator in power. So, why do nondemocratic leaders pursue
costly ideological goals abroad rather than efficiently use those
resources for payoffs at home? I argue that ideological dictators
use revisionist demands to send signals of ideological commitment to
their most important domestic supporters. This article demonstrates
through an analysis of dyadic MIDS and a monadic analysis of the
first use of violent force that it is rational signaling behavior,
and not the personal ideological preferences of leaders, which links
domestic revisionist ideology to international revisionist
conflict.}
}
For attribution, please cite this work as:
Hanna, Tom. n.d. “Theory: Brief Outline, Tables, and
Graphs.”
