Tsebelis G. Veto Players: How Political Institutions Work

Подождите немного. Документ загружается.

351

book. In fact, Proposition 3 is testing not simply the number of veto players but also their

ideological distances (see Proposition 1.4 of Chapter 1).

To test their predictions Keefer and Stasavage use a dataset including 78 countries over

20 years (for the 1975-94 period). Their dependent variable is inflation. Their independent

variables include central bank independence and a series of institutional variables. They use as

indicators of central bank independence both the “legal independence” measuring a series of

institutional indicators (see previous section) and turnover rates of the governors of the central

bank (a variable that is considered in the literature to reflect better the actual independence in

developing countries). Their institutional variables can be found in a database on political

institutions assembled by Beck, Clarke, Groff, Keefer and Walsh (1999). The variable “checks”

is “based on a formula which first counts the number of veto players, based on whether the

executive and legislative chamber(s) are controlled by different parties in presidential systems

and on the number of parties in the government coalition for parliamentary systems” (Keefer and

Stasavage (2000a: 19)). Polarization is measured “according to whether data sources indicated

parties (the four biggest ones) as having an economic orientation that was left, center or right…

the maximum difference between those entities that comprise the checks indicator explained

earlier. This maximum constitutes the political polarization measure” (Keefer and Stasavage

(2000a: 20)).

The reader can verify that the method used for the identification of institutional variables

is quite closely connected to the methods used in this book. The biggest difference is that

ideological positions of different veto players are not accurately identified, which is compensated

by their sample size. Keefer and Stasavage use interactive variables (CBI*veto players) in order

to test whether institutions matter. In their analysis they find that central bank independence has

352

no effect on inflation when it enters linearly (a result reported quite frequently in the literature),

but the coefficients of the interactive terms are negative and significant. They conclude: “More

concretely, in a parliamentary system with three party governing coalition… a one standard

deviation increase in legal central bank independence would be predicted to reduce annual rate of

inflation by approximately 20 percent. In contrast, in a parliamentary system with a single party

majority…the predicted change in inflation would be close to zero… This suggests an

explanation for Cukierman, Webb, and Neyapti’s finding that CBI is significantly and negatively

correlated with inflation in advanced industrial countries but not in developing countries:

developing countries, on average, have lower levels of checks and balances.” ((Keefer and

Stasavage (2000a: 23)). When the variable “political polarization” is included in the regressions

the results indicate that “checks and balances make the biggest contribution to central bank

effectiveness in more polarized societies” (Keefer and Stasavage (2000a: 33)).

CONCLUSIONS

In this chapter I developed a simple model, which generates the expectation that for any

given legislation, the independence of bureaucrats and the judiciary will be positively related to

the veto players (number and distances among them) that control the legislative process. This

model was then tested with existing data on judicial and bureaucratic independence.

At the theoretical level I made the distinction between countries with or without

constitutional courts. I argued that constitutional courts are additional veto players (since for all

practical purposes they cannot be legislatively overruled). However, because of the rules of

selection of these courts most of the time they are absorbed as veto players by the existing

political ones. The empirical findings about judges were based on four different indexes of

353

judicial independence generated by combinations of some 20 advanced industrialized countries,

and corroborated the expectations of the model of the first section.

With respect to bureaucracies I made the distinction between institutional and behavioral

independence in order to differentiate the argument presented in this book from the standard

arguments in the literature. I was able to identify a case (central bank activities) where the

expectations generated from my argument could be tested. Examining empirical evidence I

found out that central bank independence is correlated with both veto players and federalism. I

argued that most likely CBI is an institutional variable and as a consequence the theory presented

in this book cannot account for the reasons that central banks are more independent in countries

with many veto players. However, focusing on behavioral independence (in the fight against

inflation) I presented evidence that the expectations generated in the first part of this chapter are

strongly corroborated by a dataset covering a large number of countries.

354

TABLE 10.1

Data on Judicial Independence

COUNTRY JUDPOL

JUDREV

EXPERTS

STRICT

LIABILITY

DEC POL

veto

players DURATION

Australia 3

2.33

3 1 2 9

Austria 3 3 2

3 2 2 8

Belgium 3 3

3.5

3 1 3 4.8

Canada 3

2.33

3 1 1 8

Denmark 1 2 1

1 1 3 8

Finland 1 1 1

1 1 3 4.8

France 4 3

3.7

1 2 2 6.8

Germany 4 4

3.46

3 2 3 6.3

Greece 2 2 1 2 1

Iceland 1 2 1 1 3

Ireland 2 2 1

1 2 2 6.4

Israel 1

4.5

1 2 2 2.4

Italy 4 3

3.33

2 2 3 1.3

Japan 2

2.17

1 1 1 9.4

Luxembourg 1 1 3

1 1 2 4

New Zealand 1

1 1 1 6

Norway 1 2 1

1 2 2 4

Portugal 3 2 1 2 2

Spain 3 3

3 2 1 6

Sweden 2 2

2.5

1 2 2 4.2

Switzerland 2 1 3 1 3

Netherlands 2 1

4.2

1 1 3 2.5

United Kingdom 2 1

2.1

2 2 1 8

US 4

4.42

3 1 3 6.9

355

TABLE 10.2

Correlations between independent variables with and without Australia, Canada, Israel,

and the Netherlands

10.2A 10.2B

Judrev experts Strict

liab.

judpol judrev experts Strict

liab.

judpol

Judrev

(Lijphart)

1.000 1.000

Experts

.1261 1.000 .6446 1.000

Strict

liability

.4003 .7912 1.000 .5286 .9132 1.000

Judpol

(Alivizatos)

.8660 .3259 .5547 1.000 .8346 .7603 .6471 1.000

Correlations among different indexes are similar whether the four countries are included

or not, with few exceptions (noted in bold)

356

TABLE 10.3

Judicial Independence as a function of veto players

Independent

variable

Judpol

(Alivizatos)

Judrev

(Lijphart)

Strict

liability

Strict

liability

Strict

liability

Experts*** Experts*** Experts*** Judpol***

(Alivizatos)

veto players

(z or t)

1.292004**

(2.049)

.3405817

(1.060)

.6194123**

(1.686)

.6563661**

(1.779)

.5765635*

(1.454)

.8774138***

(4.415)

.7278529***

(3.107)

.8310000***

(4.297)

.6307947**

(2.095)

Dependent

variables

Polarization

(z or t)

3.259719***

(3.067)

.8772162**

(1.730)

-.5285295

(-0.954)

- - -.1133333

(-0.340)

- - 1.674503***

(3.703)

Decentralization

(z or t)

.9234209***

(2.543)

.9548024***

(3.172)

.3997773*

(1.360)

.3580203

(1.248)

- -.232069

(-1.279)

- - .5092715***

(2.626)

Duration

(z or t)

- - - - -.0469358

(-0.370)

- -.06612

(-0.807)

- -

#obs. 18 24 20 20 20 14 14 14 18

PseudoR

Adjust. R

0.3712 0.2291 0.1389 0.1167 0.0818 0.5643 0.5605 0.5733 0.5565

-Columns with the independent variable marked by *** use OLS estimations (t-statistic and adjusted R

-statistical significance determined by one-tailed tests (* significant at the .10, ** at the .05, and *** at the .01 level)

TABLE 10.4

Data on Central Bank Independence (from Bernhard (1998))

COUNTRY DEC

veto

players

Grilli et.

al. (1991)

Alesina

Summers

(1993)

Cukierman

(1991)

Total

Australia 3 2 0.6 0.5 0.31 0.47

Austria 3 2 0.6 0.625 0.58 0.6

Belgium 3 3 0.47 0.5 0.19 0.39

Canada 3 1 0.73 0.625 0.46 0.61

Denmark 1 3 0.53 0.625 0.47 0.54

France 1 2 0.47 0.5 0.28 0.42

Germany 3 3 0.87 1 0.66 0.84

Ireland 1 2 0.47 0.625 0.39 0.49

Italy 2 3 0.33 0.45 0.16 0.33

Japan 1 1 0.4 0.625 0.16 0.4

New Zealand 1 1 0.2 0.25 0.27 0.2

Norway 1 2 0.44 0.5 0.14 0.4

Spain 3 1 0.33 0.375 0.21 0.31

Sweden 1 2 0.44 0.5 0.27 0.4

Switzerland 3 3 0.8 1 0.68 0.83

Netherlands 1 3 0.67 0.625 0.42 0.57

United Kingdom 2 1 0.4 0.5 0.31 0.42

US 3 3 0.8 0.875 0.51 0.73

358

TABLE 10.5

Different measures of Central Bank Independence (CBI) as a function of veto players and

decentralization

CBI Grilli et.

al.(1991)

Alesina +

Summers

(1993)

Cukierman

(1991)

Total CBI

veto players

(t)

.0998795

(2.361)

.1175301**

(2.451)

.0703614*

(1.563)

.0976386

(2.379)

DECENTRALIZATION

(t)

.0862651

(2.377)

.0634337*

(1.542)

.0662048*

(1.715)

.0727952

(2.067)

N 18 18

Adj-R

0.3958 0.3115

0.2018

0.3612

359

FIGURE 10.1

BY FIRST MOVER (BUREAUCRACY OR JUDICIARY)

First mover outside core (J or K) selects

closest point inside core (J' or K');

First mover inside core (L1 or L2) selects

own ideal point.

360

CHAPTER 11: VETO PLAYER ANALYSIS OF EU INSTITUTIONS

The European Community fascinates observers and scholars because it is a unique object

of study. Because the institutional structure of the EU is new, a series of neologisms have been

invented to describe it. It is "neither a state nor an international organization" (Sbragia (1992:

257)); "less than a Federation, more than a Regime" (W. Wallace (1983: 403)); "stuck between

sovereignty and integration" (W. Wallace (1982: 67)); a "part formed political system." (H.

Wallace (1989: 205); "institutionalized Intergovernmentalism in a supranational organization"

(Cameron (1992: 66)); the "middle ground between the cooperation of existing nations and the

breaking of a new one" (Scharpf (1988: 242)). Some scholars have even seen advantages in the

situation: Sbragia (1992: 258) approvingly quotes Krislov, Ehlermann, and Weiler claiming:

"The absence of a clear model, for one thing, makes ad hoc analogies more appropriate and

justifiable. If one may not specify what are clear analogies, less clear ones may be appropriate."

In this chapter instead of using (appropriate or inappropriate) analogies, I apply the veto

players theory and examine the logic and the outcomes of decision making generated by the

different legislative procedures adopted in successive treaties, and compare my conclusions with

other institutional analyses. As a result, this chapter has three important characteristics: first, it

studies several different institutional structures that prevail in the same territory; second, it

advances the application of the veto players theory, because the EU institutions are quite

complex: they involve three legislative institutions, each one of them deciding by different

formal or actual majorities, and the right to set the agenda sometimes includes restrictions,

sometimes is shared, and sometimes shifts among actors; third, because the study of EU

institutions and policies has progressed further than case studies, the predictions of different

theories are sharper, and the data collected enable comparisons on relatively solid basis. In short,

this chapter will make comparisons across different institutional settings (that involve the same

geographic area), will push theory to cover more complicated institutions, and will use empirical

tests to cooroborate more detailed empirical predictions than any of the previous chapters.