Dimensions of Governance: A Selectorate Pilot Study Bruce Bueno de Mesquita Alastair Smith New York University.

Dimensions of Governance: A Selectorate Pilot Study

Bruce Bueno de Mesquita

Alastair Smith

New York University

2

Objectives Measure selectorate theory governance variables

• Selectorate Size (S)• Nominal

• Real

• Blocs

• Winning Coalition Size (W)• Absolute

• Relative

• Blocs Demonstrate the new data’s potential to improve the

government’s ability to identify, explain, and alter:• Prospects for democratization

• Emerging national security risks

3

Pilot Project Countries: Expert Survey

China DPRK Egypt Iran Nigeria Pakistan Russia Saudi Arabia

Syria Venezuela

Experts Asked

7 5 22 14 14 15 8 12 5 11

Complete Surveys

5 5 6 6 4 4 5 4 5 5

Usable Data

5 4 6 5 4 4 5 4 5 5

4

Standard View of Governance

Autocracy Continuum Democracy

Alternative Arbitrary

Dividing Lines Between Systems

Autocracy Democracy

5

Limitations of Standard View Dimensions of democracy are conceived as linear and additive Theoretical links to governance or national security are unclear

• Institutions

• Behavior

• Judgements Measures over-weight elections, even if they are rigged,

relative to:• Accountability

• Transparency

• Bloc-controlled choices vs. individual choices regarding

• Leaders

• Policies

• Institutions

6

Governance Dimensions: Nested Selectorate Institutions

Population

Nominal Selectorate

Real Selectorate

Winning Coalition

Blocs

7

Ten Countries in 2006: W and S

010

2030

4050

mea

nWag

e

0 10 20 30 40 50 60 70 80 90 100meanNSpop

Co

alit

ion

Siz

e a

s %

of

Po

pu

latio

n

Nominal Selectorate as % of Population

8

Revenue-Expenditure

PublicGoods

PrivateGoods

Leader’sDiscretionary

Funds

The coalition’swelfare dependson W:

Size of W

Totalexpendituresincrease with Wand decrease asS increases

Tax ratesdecrease as Wincreases andas S decreases

Opportunities forkleptocracy dependson W:

Increases as W increases

Increases as Wdecreases

0

Size of W

Size of W

Leisuredecreases andLabor increasesas W increasesand as Sdecreases

Selectorate View of Policy Choices

Source: Bueno de Mesquita, Smith, Siverson and Morrow, The Logic of Political Survival. Cambridge: MIT Press, 2003

9

Selectorate View: Democratization Prospects

0 Coalition Size (W) 1000

Predicted Diminution of Coalition Size/Democracy

Predicted Expansion of

Coalition Size/ Democracy

Low

High

Leader’s Welfare

Predict No Change in Coalition

Size

Predict No Change in Coalition

Size

Credible Threat to Political Institutions Institutions Secure: No Incentive for Anti-Government

Movement

Institutions Secure: Too

Low Probability of Successful

Anti-Government Movement

Contract W

Expand W

Blue line = Predicted response to threats to the regime in power. Source: Bueno de Mesquita and Smith. 2007. “Political Survival and Endogenous Institutional Change.” NYU Working Paper.

10

Democratization

Predicting Changes in Governance

11

Predicting Polity’s Estimate of Democracy-Autocracy (t+5 Years):

A Difficult TestDependent Variable: Polity (t+5 years) Given a Credible Threat to Regime Survival

Polity-Based Model Selectorate Model

Polity -22.28 (5.68) 0.000 %Nominal Selectorate (NS) -0.11 (0.04) 0.002 S/NS 0.06 (0.02) 0.003 W -4.26 (2.05) 0.037 W/S 0.41 (0.19) 0.030 Institutional Clarity -5.97 (2.73) 0.029 Size of Ethnic Bloc in W 10.33 (2.43) 0.000 Size of Religious Bloc in W -10.02 (2.34) 0.000 Size of Military Bloc in W -0.10 (0.04) 0.007 Size of Other Blocs in W -0.19 (0.05) 0.000 Oil 0.005 (0.02) 0.807 0.06 (0.04) 0.120 Polity*Oil -0.02 (0.07) 0.735 W*Oil 0.001 (0.001) 0.882 Aid (%GDP) -0.19 (0.08) 0.016 -0.98 (0.38) 0.010 Polity*Aid 0.79 (0.58) 0.177 W*Aid -0.65 (0.22) 0.004 Log(Per Capita Income) -1.93 (0.34) 0.000 14.33 (4.35) 0.001 Polity*Log(Per Capita Income)

3.61 (0.79) 0.000

W*Log(Per Capita Income) 0.51 (0.24) 0.032 Log(Population) -0.05 (0.17) 0.758 8.47 (2.79) 0.002 Year -0.002 (0.02) 0.909 -0.38 (0.13) 0.003 N=95 χ2 = 195.33

p<0.000 Pseudo R2 = 0.50

N=95 χ2 = 279.46 p<0.000 Pseudo R2 = 0.72

Significance of Selectorate model’s explanatory improvement : p < 0.000

12

Predicting Governance in Five Years

.24

.26

.28

.3

2005 2006 2007 2008 2009 2010 2011Future

Win5 Polityin5

Pre

dic

ted

Co

alit

ion

Siz

e a

nd

Po

lity

Sco

reEgypt

Year

The selectorate model is optimistic about Egypt; the Polity model is pessimistic

13


.14

.16

.18

.2.2

2.2

4

2005 2006 2007 2008 2009 2010 2011Future

Win5 Polityin5

Pre

dic

ted

Co

alit

ion

Siz

e a

nd

Po

lity

Sco

reChina

Year

Polity model was more optimistic about China but is now becoming pessimistic relative to the selectorate model

14


.75

.8.8

5.9

.95

2005 2006 2007 2008 2009 2010 2011Future

Win5 Polityin5

Pre

dic

ted

Co

alit

ion

Siz

e a

nd

Po

lity

Sco

reVenezuela

Year

The selectorate and Polity models anticipate improvement followed by decline in Venezuela

15


.1.2

.3.4

.5.6

2005 2006 2007 2008 2009 2010 2011Future

Win5 Polityin5

Pre

dic

ted

Co

alit

ion

Siz

e a

nd

Po

lity

Sco

reNigeria

Year

The selectorate model anticipates a more precipitous decline for Nigeria than does Polity

16


.3.4

.5.6

.7

2005 2006 2007 2008 2009 2010 2011Future

Win5 Polityin5

Pre

dic

ted

Co

alit

ion

Siz

e a

nd

Po

lity

Sco

rePakistan

Year

The models share a common view for Pakistan’s governance future

17


.4.4

5.5

.55

.6

2005 2006 2007 2008 2009 2010 2011Future

Win5 Polityin5

Pre

dic

ted

Co

alit

ion

Siz

e a

nd

Po

lity

Sco

reIran

Year

The selectorate view anticipates improvement followed by new erosion in democracy in Iran

18


-.05

0.0

5.1

2005 2006 2007 2008 2009 2010 2011Future

Win5 Polityin5

Pre

dic

ted

Co

alit

ion

Siz

e a

nd

Po

lity

Sco

reSaudi Arabia

Year

Both models anticipate very poor governance performance for Saudi Arabia through 2011

19


.1.1

5.2

.25

2005 2006 2007 2008 2009 2010 2011Future

Win5 Polityin5

Pre

dic

ted

Co

alit

ion

Siz

e a

nd

Po

lity

Sco

reSyria

Year

Both models anticipate improvement in Syria around 2010, but the selectorate takes a dimmer view of current Syrian governance than does Polity

20


.5.6

.7.8

.9

2005 2006 2007 2008 2009 2010 2011Future

Win5 Polityin5

Pre

dic

ted

Co

alit

ion

Siz

e a

nd

Po

lity

Sco

reRussia

Year

The selectorate view expects russia to get worse before it gets better (around 2010)

21

A Sample Application in the National Security Arena

Predicting Coups, Civil War,

and Interstate War

22

Predicting Coups, Civil War or Interstate War in One Year

Conflict (t+1 year) Conflict (t+1 year) Polity Democracy-Autocracy 8.58 (3.20) 0.007 S/NS 0.02 (0.01) 0.040 W 1.00 (0.26) 0.000 W/S -0.07 (0.03) 0.005 Clarity or Opacity -1.73 (0.44) 0.000 Aid/GDP 0.06 (0.06) 0.342 0.14 (0.07) 0.054 Log(Income per Capita) -0.54 (0.22) 0.012 -0.31 (0.27) 0.239 Polity*Log(Income per Capita)

-1.23 (0.46) 0.008

W*Log(Income per Capita) -0.12 (0.03) 0.000 Log(Population) -0.67 (0.19) 0.000 -0.83 (0.24) 0.001 Credible Threat -0.50 (0.20) 0.014 -0.31 (0.18) 0.074 Constant 14.04 (4.30) 0.001 14.46 (5.05) 0.004 N=251 χ2 = 59.23 PRE = 12% N=260 χ2 = 72.50 PRE = 24%

23

Nominal Selectorate

Survey Question

How many people have at least a nominal say in choosing the national leader whether these people are in the selectorate or not?

(Recall, for instance, all adult citizens in the Soviet Union had a nominal say through the ballot box even if they were not members of the Communist Party; that is, even if they were not members of the selectorate.)

24

Nominal Selectorate (0-100%)


Syria Venezuela

Mean 3.6 71.3 49.3 38.5 32.0 37.7 73.2 33.3 100 55.0

Median 3.6 100 40.1 41.5 31.3 29.7 100 0.0 100 46.6

25

Middle Eastern Nominal Selectorates

02

04

06

08

01

00

1955 1965 1975 1985 1995 2005Year

IranNS SaudiNSEgyptNS SyriaNS

No

min

al S

ele

cto

rate

as

%P

op

ula

tion

26

Asian Nominal Selectorates

02

04

06

08

01

00

1955 1965 1975 1985 1995 2005Year

ChinaNS DPRKNSPakistanNS

No

min

al S

ele

cto

rate

as

%P

op

ula

tion

27

Oil Exporters on Three Continents: Nominal Selectorates

02

04

06

08

01

00

1955 1965 1975 1985 1995 2005Year

VenezuelaNS RussiaNSNigeriaNS

No

min

al S

ele

cto

rate

as

%P

op

ula

tion

28

Selectorate Size

Survey Question

What was the size of the selectorate as a percentage of the total population?

(Recall that the selectorate is the politically relevant class of the society, those with the opportunity to influence the selection of leaders.)

29

Selectorate: % of Nominal Selectorate Selectorate: % of Population


Syria Venezuela

S to Nominal S: Mean

0.1 0.5 15.9 40.4 86.5 48.0 32.4 0.0 9.0 93.3

S to Nominal

S: Median

0.1 0.1 9.9 46.1 87.2 40.4 33.5 100 2.7 94.1

S: Mean 0.0 0.4 7.4 15.0 25.6 19.4 30.3 0.0 9.0 51.0

S: Median 0.1 0.1 3.9 8.5 22.1 8.5 33.5 0.0 2.7 43.6

30

Middle Eastern Ratio of Selectorate to Nominal Selectorate

02

04

06

08

01

00

1955 1965 1975 1985 1995 2005Year

SaudiStoNS EgyptStoNSIranStoNS SyriaStoNS

Se

lect

ora

te a

s %

No

min

al S

ele

cto

rate

31

Asian Ratio of Selectorate to Nominal Selectorate

02

04

06

08

01

00

1955 1965 1975 1985 1995 2005Year

ChinaStoNS PakistanStoNSDPRKStoNS

Se

lect

ora

te a

s %

No

min

al S

ele

cto

rate

32

Oil Exporters on Three Continents: Selectorate to Nominal Selectorate

02

04

06

08

01

00

1955 1965 1975 1985 1995 2005Year

VenezuelaStoNS RussiaStoNSNigeriaStoNS

Se

lect

ora

te a

s %

No

min

al S

ele

cto

rate

33

Winning Coalition Size

Survey QuestionWhat was the size of the winning coalition?

(Recall that this is not based on the total number of people who supported the leader; it is the number of supporters who are essential

to maintaining the leader in power and who receive policy concessions and/or personal

benefits for their support).

34

Winning Coalition: % of Selectorate Winning Coalition: % of Population


Syria Venezuela

W/S: Mean

0.3 0.2 41.8 6.1 14.9 14.9 25.8 3.8 2.1 27.0

W/S: Median

0.3 0.2 39.8 6.0 0.3 17.8 27.8 2.4 0.6 28.7

W: Mean 0 0 3.1 1.0 7.8 3.2 8.7 0 0.7 14.4

W: Median 0 0 3.1 1.4 0 3.3 0.9 0 0 15.2

35

Middle Eastern Coalition to Selectorate Size (W/S)

020

4060

8010

0

1955 1965 1975 1985 1995 2005Year

IranWtoS SaudiWtoSEgyptWtoS SyriaWtoS

Co

alit

ion

as

% S

ele

cto

rate

36

Asia: Coalition Size to Selectorate Size (W/S)

020

4060

8010

0

1955 1965 1975 1985 1995 2005Year

ChinaWtoS DPRKWtoSPakistanWtoS

Co

alit

ion

as

% S

ele

cto

rate

37

Oil Exporters on Three Continents: Coalition to Selectorate Size (W/S)

020

4060

8010

0

1955 1965 1975 1985 1995 2005Year

VenezuelaWtoS RussiaWtoSNigeriaWtoS

Co

alit

ion

as

% S

ele

cto

rate

38

Bloc Size as % of Coalition


Syria Venezuela

Mean 100 100 100 48.2 100 19.4 0 100 93.3 96.7

Median 100 100 100 0.3 100 0.1 0 100 100 100

39

Middle East: Blocs as % of Coalition Size

020

4060

8010

0

1955 1965 1975 1985 1995 2005Year

IranBloc SaudiBlocEgyptBloc SyriaBloc

Blo

cs a

s %

Win

nin

g C

oa

litio

n

40

Asia: Blocs as % of Coalition Size

020

4060

8010

0

1955 1965 1975 1985 1995 2005Year

ChinaBloc DPRKBlocPakistanBloc

Blo

cs a

s %

Win

nin

g C

oa

litio

n

41

Oil Exporters on Three Continents: Blocs as % of Coalition Size

020

4060

8010

0

1955 1965 1975 1985 1995 2005Year

VenezuelaBloc RussiaBlocNigeriaBloc

Blo

cs a

s %

Win

nin

g C

oa

litio

n

42

Comparing Polity Scores and Coalition Size

43

Comparing Coalition Size to Polity’s Democracy-Autocracy

02

.00

e-0

64.0

0e-0

66.0

0e-0

68.0

0e-0

6.00

001

Coa

litio

n

51

5P

olit

y

1955 1965 1975 1985 1995 2005Year...

Polity Coalition

China Coalition Size

44


0.0

00

5.0

01

.00

15

Coa

litio

n

51

5P

olit

y

1955 1965 1975 1985 1995 2005Year...

Polity Coalition

DPRK Coalition Size

45


0.0

00

2.0

00

4.0

00

6.0

00

8.0

01

Coa

litio

n

15

20

Po

lity

1955 1965 1975 1985 1995 2005Year...

Polity Coalition

Egypt Coalition Size

46


0.0

05

.01

.01

5C

oa

litio

n

06

5P

olit

y

1955 1965 1975 1985 1995 2005Year...

Polity Coalition

Iran Coalition Size

47


05

10

15

20

Coa

litio

n

15

90

Po

lity

1955 1965 1975 1985 1995 2005Year...

Polity Coalition

Nigeria Coalition Size

48


.00

05

.00

1.0

01

5.0

02

Coa

litio

n

15

90

Po

lity

1955 1965 1975 1985 1995 2005Year...

Polity Coalition

Pakistan Coalition Size

49


01

02

03

0C

oa

litio

n

70

85

Po

lity

1955 1965 1975 1985 1995 2005Year...

Polity Coalition

Russia Coalition Size

50


0.0

00

2.0

00

4.0

00

6.0

00

8C

oa

litio

n

05

10

15

20

25

Po

lity

1955 1965 1975 1985 1995 2005Year...

Polity Coalition

Saudi Arabia Coalition Size

51


02

46

8C

oa

litio

n

58

5P

olit

y

1955 1965 1975 1985 1995 2005Year...

Polity Coalition

Syria Coalition Size

52


05

10

15

20

25

Coa

litio

n

35

95

Po

lity

1955 1965 1975 1985 1995 2005Year...

Polity Coalition

Venezuela Coalition Size

53

Future Work The pilot study sows it is feasible to collect subtle data on numerous

governance dimensions suggested by the logic of the selectorate theory.

The selectorate dimensions seem to be better predictors of democratization, defined in Polity terms or in selectorate terms (not shown here), than are Polity scores themselves.

The selectorate dimensions seem to better predict conflict than Polity. The pilot study was limited to 10 non-randomly selected countries

chosen by the contractor. The evidence encourages the view that data on selectorate

dimensions should be collected for all countries for at least the past 50 years and should be updated annually.

The evidence suggests that it may make sense to commission studies of general patterns of governance and of specific countries and their susceptibility to democratization in the coming years.

Dimensions of Governance: A Selectorate Pilot Study Bruce Bueno de Mesquita Alastair Smith New York University.

Documents

governance slide

selectorate model slide

predicting governance

polity slide

s slide

pessimistic slide

iran slide

venezuela slide