Party Creation as an Autocratic Survival Strategy



Download 201.26 Kb.
Page4/4
Date conversion21.03.2016
Size201.26 Kb.
1   2   3   4

*Based on 40 percent of sample.


The full sample is made up of 128 regimes in which the dictator or ruling group created a party after the seizure of power. It excludes monarchies and regimes in which the dictator coopted and made use of a preexisting party as well as those in which the ruling group was organized as a party prior to achieving office.
Table 2: The Effect of Party Creation on the Likelihood of Coups

Hazard Model: Cox Regression







Banks Coup Data

Banks

Coup Data



Belkin Coup Data

Belkin Coup Data

Banks with Frailty Terms

Belkin with Frailty Terms


Created Party


0.35

(.001)


0.31

(.009)


0.52

(.005)


0.45

(.011)


0.25

(.002)


0.45

(.007)


Monarchy


0.01

(.000)


0.04

(.005)


0.07

(.000)


0.13

(.002)


0.00

(.000)


0.05

(.000)



Growth


0.98

(.144)


1.00

(.987)


0.99

(.553)


1.00

(.818)


0.97

(.075)


0.99

(.324)



Log GDP


0.77

(.151)


0.77

(.284)


1.00

(.970)


1.11

(.497)


0.75

(.347)


0.88

(.545)



Civil War


1.81

(.074)


1.14

(.817)


1.30

(.338)


1.35

(.450)


1.83

(.199)


0.61

(.225)



Age of Leader


1.00

(.885)


1.02

(.284)


(0.99)

(.187)


1.00

(.789)


0.99

(.604)


0.97

(0.05)


Couprisk


2.37

(.000)


2.07

(.000)


1.96

(.000)


2.19

(.000)


2.93

(.000)


2.00

(.000)



Counter-balancing





0.76

(.064)





0.82

(.039)








N

1276

611

1323

611

1276

1323

Table 3: Effect of Party Creation on the Likelihood of Coups


Logit





Belkin/Schofar Coup Data


Banks Coup Data


Combined Coup Data


Combined Coup Data

Created Party


-0.51

(.064)


-0.79

(.029)


-0.64

(.015)



-0.68

(.073)


Monarchy


-1.81

(.003)



-3.05

(.006)


-2.03

(.001)


-1.25

(.126)


Growth


-0.01

(.373)


-0.05

(.013)


-0.04

(.008)



-0.02

(.367)


Log GDP


-0.23

(.081)


-0.34

(.062)


-0.28

(.028)



-0.21

(.240)


Civil War


-0.06

(.839)


0.16

(.662)


0.17

(.540)



-0.64

(.185)


Year


-0.06

(.000)


-0.05

(.010)


-0.05

(.000)



-0.09

(.015)


Coup Risk


0.24

(.047)


0.14

(.359)


0.23

(.038)



0.29

(.084)


Regime Duration

-0.01

(.671)



0.01

(.556)


-0.00

(.939)


-0.01

(.669)


Counter-balancing











-0.06

(.643)


N

1161

1118

1161

550

Log likelihood

-286.6

-193.8

-305.0

-159.9

Pseudo R2

0.13

0.12

0.15

0.15

Table 4: Predictors of Parties Created Post-Seizure


Logit
Dependent Variable: Created Party





Coefficient


Coup Risk


-0.20

(.030)



Monarchy


-14.74

(.000)



Growth


0.08

(.183)



Log GDP


-0.84

(.000)



Civil War


0.31

(.183)



Year


-0.03

(.002)



Leader’s Age


-0.02

(.054)



Regime Duration

0.29

(.000)



N

1159

Log likelihood

-369.3

Pseudo R2

0.54

Table 5:

Effect of Instrumented Party Creation on the Likelihood of Coups
Logit





Belkin/Schofar Coup Data


Banks Coup Data


Combined Coup Data


Instrumented Created Party


-1.24

(.070)


-0.75

(.398)


-1.20

(.068)



Monarchy


-2.42

(.006)



-3.04

(.028)


-2.49

(.004)


Growth


-0.02

(.331)


-0.04

(.021)


-0.04

(.006)



Log GDP


-0.23

(.081)


-0.34

(.135)


-0.38

(.018)



Civil War


-0.04

(.900)


0.20

(.586)


0.18

(.527)



Year


-0.06

(.000)


-0.05

(.011)


-0.05

(.000)



Coup Risk


0.17

(.175)


0.16

(.303)


0.18

(.137)



Regime Duration

0.00

(.856)



0.01

(.618)


-0.01

(.777)


N

1159

1116

1159

Log likelihood

-283.1

-193.6

-302.5

Pseudo R2

0.13

0.10

0.14

Table 6: The Effect of Military Professionalization* on Post-Seizure Party Creation


Party Creation Strategy




Military Type


Party Created

(%)


Allied with Preexisting Party

(%)



No Parties

(%)


N

Recently Indigenized


58

21

21

19

Professionalized

33

9

58

57

* Indigenization of the officer corps and creation of a new military force by foreign occupiers within 20 years of the seizure of power are used as proxies for lack of professionalization.


Based on partial sample
Figure 3: Divide the Spoils Game with Military Factions



1 Thanks to Ellen Lust-Okar and Carles Boix for comments on earlier versions of this paper, to Joe Wright for help finding data, to Erica Frantz for research assistance, and to John Zaller for many kinds of help.

2 See, for example, Chehabi and Linz (1998)

3 Data collected by the author.

4The next three words in Downs’s definition are “by legal means” so he intended his definition to be limited to electoral parties, not parties organized to seize power by any means necessary.

5 Other standard theories about why parties exist, such as those by Aldrich (1995) and Cox and Mc Cubbins (1993) have little relevance for the study of modern authoritarian parties since they posit party development within legislatures governed by majority rule. Parties in some monarchies have developed in this way, but parties that have seized power have been formed by exactly those citizens excluded from representation in legislatures (if they exist in the ancien regime). Parties created by dictators after seizures of power build an organizational structure for the dictator’s supporters regardless of the institution they inhabit, often rely heavily on public employees initially, and usually predate the creation of a legislature.

6 See DeNardo (1985) for a detailed elaboration of the dilemma facing the leaders of revolutionary movements.

7 Referring to this form of power sharing as delegation implies that the dictator’s own interests are served by it. That seems a dubious assumption since presidents’ terms are more likely to be limited in such systems. My observation suggests that presidents agree to such arrangements when power is relatively dispersed among elite factions within the ruling group, which forces the president to grant this form of power sharing. If power is already concentrated in the president’s hands when the party is created, he has little reason to delegate powers that may increase regime survival but not his own time in office.

8 Data from Belkin and Schofer (2003). My thanks to them for sharing their data.

9 Monarchs sometimes allow parties to develop and then form alliances with one of them, but they rarely create the kind of party discussed here, the hegemonic support organization of the regime. Instead, if they allow the existence of parties, they remain above the fray and play the parties off against each other.

10 See Fontana (1987) for a detailed description of the months of negotiations before the Argentine coup of 1976 and the military’s later inability to enforce the bargain it had negotiated.

11 The reasoning here is analogous to Barron and Ferejohn (1989). In divide-the-dollar games, the only cost to the agenda setter of being replaced is losing agenda setting power, but dictators face the risk of much higher costs.

12 Officers’ concern about unity, cohesion, and professional efficacy has been noted by military sociologists going back to Janowitz (1960). O’Donnell, Schmitter, and Whitehead (1986) famously claimed that all transitions to democracy began with splits within the authoritarian ruling group. This claim may not be generally true, but it is true for most transitions from professionalized military rule.

13 See Geddes (2003) for a description of this dataset and the coding rules used to produce it.

14 As Magaloni (2006) has observed, authoritarian governments seek to be elected by supermajorities in high turnout elections because such outcomes serve as effective deterrents to elite defections, while bare majorities would not. A variation on minimum-winning-coalition logic operates within the regime elite, but not in elections or legislatures. Dictators can afford to exclude some erstwhile elite supporters from spoils as long as they can deter rivals from challenging them because no one can do better by joining the opposition. In a dictatorship, joining the opposition simply means excluding oneself from the only game in town and any benefits that access to government might deliver.



1   2   3   4


The database is protected by copyright ©essaydocs.org 2016
send message

    Main page