Tsebelis G. Veto Players: How Political Institutions Work

Подождите немного. Документ загружается.

341

Strom, who explicitly compares the two systems and addresses the veto players argument, so I

will focus on his presentation and address the question of bureaucratic independence on the basis

of his analysis.

According to Strom the major characteristic of a parliamentary regime is the simplicity of

its structure of delegation. Indeed, from elections, to selection of prime minister, to the selection

of ministers, to the instructions to bureaucrats the whole political life is a series of delegations

from individual or collective principals to individual or collective agents. By contrast, delegation

in presidential systems ignoring the issue of individual or collective actors, takes place from

single principles to multiple agents (the people select multiple institutions), from multiple

principals to single actors (the different legislative institutions delegate to single agencies, or

from multiple principals to multiple agents (sometimes the legislative institutions may have

different agencies compete for some particular task).

However, this is the ideal type representation of the two systems, and Strom recognizes

that reality may come in different shades (in fact, he acknowledges that some of these shades

determined by the veto players theory). I want to expand on these points, and argue that the

single chain of delegation is a simplification that may be important in order to highlight

characteristics of presidentialism like the explicit supervision of bureaucracies by the different

principals, compared to the absence of similar hearings in parliamentary systems, but ignores

decisionmaking inside the government who in case of multiparty governments may be a single

principal only in a very abstract and black box-like way.

Let us try to open the black box of government: according to Strom the prime minister

delegates to the ministers, and the ministers delegate to the bureaucrats, so the bureaucrats are

342

part of the single chain of command. So, in the idealized version of Strom’s argument, the prime

minister ultimately determines the behavior of bureaucrats.

Let us compare this point of view with the Laver and Shepsle (1997) argument of

exclusive ministerial jurisdictions discussed in Chapter 4. According to these authors, the

minister decides what the bureaucrats are going to do. Not so, Thies (2001) would claim; he

presents evidence of governments being composed of different parties not only across but also

within ministries, so that the vice-ministers come from different parties from the minister in

order to keep him in check. In Thies’ point of view the chain of command is not unambiguous.

Let us now present a different point of view presented by Mark Hallerberg and Jürgen

von Hagen (1999) who claim that with respect to budgets some countries delegate ultimate

authority to the minister of finance in order to keep the budget at the level decided by the

government. Hallerberg and von Hagen present the institutional structures of different countries

and demonstrate that in several delegations of significant policymaking powers belong to the

minister of finance. Here the chain of command involves the minister and the minister of finance

(it could involve the prime minister too).

However, as the minster of finance may play a key role on economic issues, other

ministeries may also participate in decisionmaking relevant to their jurisdictions. For example, it

does not seem plausible that in an international conference on the environment, national

delegations would include bureaucrats from the ministry of the environment without

representatives of the ministry of foreign affairs. In preparation of documents on women’s

health, bureaucrats form health and labor are likely to be involved etc.

Enter Wolfgang Mueller (2000) and his analysis of parliamentary systems as involving

parties in every step of the delegation process. In Mueller’s account parties are more present in

343

the formation of government, less in the process moving from government to individual

ministers, while interference in the delegation from ministers to civil servants is illegitimate:

“Civil servants should merely implement general rules and should do so impartially” (Mueller

2000: 311). Note that Mueller’s point here is normative (he tells us what should be happening,

not what is happening); be that as it may, it seems to dispute the delegation principle at the level

of bureaucracy altogether.

My argument is that when we try to open the black box of delegation, several plausible

theories emerge and each one of them identifies a different stream from “government” to

“bureaucrats.” It may be from the minister to bureaucrats directly (Laver and Shepsle), or it

might involve other actors prime minister (Strom), minister of finance (Hallerberg and von

Hagen), other ministers, parties etc). My theory is the only one that encompasses all these

possibilities without taking sides. It merely states that if things are important, any government

actor involved will want his point of view respected, so the outcome will be acceptable by all of

them. This is a minimalist position and does not take sides. It may be that the outcome will be

located closer to the minister, or the prime minister or of any other of these actors. More likely,

skillful bureaucrats will play each one of the principals against the others.

As we have argued in Chapter 4 it is true that presidential and parliamentary systems

differ in several dimensions (who controls the agenda, whether coalitions are fluid and address

policies or rigid and form governments). Strom’s arguments add interesting variations on the

theme of delegation, like that in parliamentary systems the ex ante selection of agents is more

effective while in presidential systems the ex post controls are institutionalized, and that

parliamentarism is characterized by more indirect delegation and accountability (since there are

additional stages involving government selection). However, focusing on bureaucracies the

344

conclusions are based on the “ideal type” (or according to Strom “maximalist”) model, which

ignores decisionmaking in government and replaces it by the principle: “civil servants have a

single principal, their respective cabinet minister” (Strom 2000: 269). If we see that bureaucrats

in coalition governments have multiple principals, then the more principals they have the more

they can play them one against the other, and the more their independence increases.

b. Empirical Evidence

A series of empirical studies have tested the deckstacking argument of McNollgast, and

found supportive evidence for it. Huber and Shipan (forthcoming), corroborated the argument

with data from labor legislation in two different instances: on the one hand American States, and

on the other European countries. They found that more veto players lead to more restrictive

legislation. Similarly, Epstein and O’Halloran (1999) corroborated the argument in several

occasions with US data. Francino (2000) analyzed EU legislation and found out more extensive

delegation to the Commission (bureaucracy) when legislation is adopted by qualified majority in

the Council than when it is adopted by unanimity.

There is then a significant amount of evidence that deckstacking occurs, and that

countries with multiple veto players have more cumbersome bureaucratic legislation. Is there any

evidence that when legislation is fixed, bureaucrats interpret it in a more independent way when

the number of veto players increases?

This is a much more difficult proposition to demonstrate, because in order to run

convincing tests one would have to include the preferences of bureaucrats as part of the analysis.

Identifying the preferences of bureaucrats in different issues imposes an almost insurmountable

problem for the analysis. However, there is one case where we can assume the preferences of

345

bureaucrats as known and see whether results approximate more or less such preferences. The

case is the question of central bank independence.

Here is my argument. Central banks have been assigned duties related to monetary

policies, exchange rates, and inflation. The literature on central banks and their significance is

vast and will not be reviewed here.

134

A series of articles in economics have measured the

independence of this particular branch of the bureaucracy. I will examine the part of this

literature that connects central bank independence (CBI) to the predictions generated by the

theory presented in this book. In particular I will examine whether central banks can exercise

more independence when there are many (and more distant) veto players.

There are two streams in the empirical literature. The first uses CBI as a dependent

variable and correlates it with institutional characteristics of countries (Bernhard (1998), Lijphart

(1999), Moser (1999), Hallerberg (2001a, 2001b)). The second correlates central bank behavior

with institutional characteristics (Lohmann (1998), Keefer and Stasavage (2000a and 2000b)).

Let me discuss each one of these streams separately.

veto players and CBI. While the papers on the effects of CBI are abundant in the

economics literature very few of them have treated CBI as a dependent variable. The

overwhelming majority look at the effects of CBI on a series of other variables, or on the

robustness of different indicators of CBI.

The existing CBI measures use some combination of institutional characteristics to assess

the independence of central banks. They involve appointment and length of tenure of the bank’s

governor, whether the bank participates in the formulation of monetary policy, whether price

stability is the major objective for the bank, and whether the bank lends money only to the

government. Positive answers to the above questions indicate higher CBI.

346

William Bernhard (1998) examined the CBI of 18 industrial democracies in the 1970-

1990 period and correlated it with a series of institutional characteristics. He used three different

indexes of CBI generated by Grili et. al. (1991), Alesina and Summers (1993) and Cukierman

(1992) but he reports results only on the basis of the average of all three. Among his independent

variables were the Alford index (an indicator of class voting), strong bicameralism, a

combination of polarization, coalition government and legislative institutions that he called

“threat of punishment.” He found that all his variables were statistically significant.

Similarly Arend Lijphart (1999) examined several different CBI measures and correlated

them with two different variables: federalism, and executive dominance. Both these variables are

correlated with veto players (see Chapter 3 which explains the positive correlation between

veto players and federalism and Chapters 4 and 7 that clarify the negative correlation with

executive dominance). He found that there is a strong correlation of CBI with federalism but no

correlation with executive dominance.

Moser (1999) created a trichotomous variable which he called “checks and balances” and

examined all OECD countries. His argument is that high checks and balances will generate

independent banks because it will be difficult for the political system to modify the charter of the

bank. He found strong corroborating results. However, Moser’s classification has been criticized

as inconsistent. For example, Hallerberg (2001b: ***) argues: “The states Moser classifies as

having strong checks and balances (Australia, Canada, Germany, Switzerland, and the United

States) are the same OECD states Lijphart (1999) classifies as truly federalist states. Yet one of

them, Canada, should not be a case of “strong checks and balances” according to Moser’s own

classification scheme, which emphasizes that chambers must have equal power and have

different procedures to elect them for the checks to be strong… More generally, on Lijphart’s

134

See Berger, et. al. (forthcoming)) for a review covering more than 150 articles.

347

federalism scale from one to five, Moser’s “strong checks and balances” states all score a five,

while the average of the remaining states is just 1.9. ” I agree with this assessments and in

addition find that other countries like Portugal, Greece, Finland, and Iceland are misclassified in

the intermediate category.

135

I will infer from all this account that Moser’s result provides some

additional evidence that federalism (but not the existence of partisan veto players) is correlated

with CBI.

Finally Hallerberg (2001b) provides a series of reasons why both overall veto players and

federalism should increase central bank independence.

136

According to his argument unitary

states with single party governments prefer both dependent central banks and flexible exchange

rates; unitary states with coalition governments prefer (moderately) independent central banks

and fixed exchange rates; federal systems with a single party government (like Canada) prefer

(moderately) independent central banks and flexible exchange rates; finally multiparty

governments in federal states prefer independent banks and flexible exchange rates. Hallerberg

finds strong empirical evidence for all his predictions, both on the independence of central banks

variable and on the exchange rate regime.

So, some of the empirical research (Bernhard 1998, Hallerberg 2001) identifies both

institutional and partisan veto players as correlated with CBI, while others find the correlation

valid only with institutional veto players (federalism). In order to see whether there is a

distinction between the two, I used Bernhard’s (1998) four CBI indexes and correlated them with

both federalism and veto players.

INSERT TABLES 10.4 AND 10.5

135

Portugal has a President with veto powers, so should be included in the countries with strong checks and

balances, while the other countries have a government which agree with the Parliament, so should be classified as

weak checks and balances.

136

See also Clark and Hallerberg (2000).

348

Table 10.5 indicates that both federalism and veto players have independently high

correlations with CBI. However, how can we interpret these correlations? All CBI indexes are

mainly institutional variables, that is, describe mainly what is written in the laws of the

corresponding countries. Nevertheless the laws are not always specific in all the questions the

coding requires. If answers to questions are provided not by laws alone but also by prevailing

practices, then the indicator is not a purely institutional measure.

Depending on whether the CBI indexes involve behavioral characteristics or are purely

institutional the interpretation of the empirical findings changes. The first possibility is that the

measures involve also behavioral characteristics. In this case, the evidence in Table 10.5

corroborates the veto players theory: the more veto players, the more independent this particular

bureaucracy. I do not have an answer to why federalism in all empirical studies has an

independent impact on CBI. One possible answer would be that in some countries (Germany,

US), states participate in the appointment of board members of the central bank. However, this is

not a universal practice.

The second possibility is that the CBI measures represent purely institutional

characteristics. In this case, CBI is not an answer to the question “if legislation is held constant,

which central banks are more independent?” which is the question addressed in the first section

of this chapter. The question pertinent to the CBI data would be “why do lawmakers in some

countries prefer higher levels of central bank independence?” and the theory presented in this

book provides no answer to this question. The answer requires a “genetic” kind of argument.

137

Hallerberg (2001b) is to my knowledge the only article providing reasons why different kinds of

governments (federal vs. unitary) and different configurations of veto players (single vs.

137

Like the one presented by Alivizatos (1995) with respect to constitutional courts (see previous section) that

ceteris paribus such courts are likely to be present in countries with serious human rights violations in their past.

349

multiple) would have difference preferences and generate different outcomes in terms of CBI

and exchange rate regimes. His argument is completely independent form the argument

presented in this book.

veto players and Central Bank behavior. Let us now move to a more precise expression of

independence: the behavior of the bank. Is this behavior affected by the political environment

under which the bank operates? In an important empirical article Susanne Lohmann (1988)

tested a series of theories related to the actual performance of the German Bundesbank. Her

conclusions were: “The behavioral independence of the German central bank fluctuates over

time with the party control of federalist veto points” (Lohmann 1988: 401).

In order to reach this conclusion Lohmann examines five competing hypotheses. They

were, first, that the central bank has full independence and is composed by technocrats; second,

that the central bank has no independence (in which case it does not matter what it is composed

of); third that the central bank has full independence and is composed by partisans; fourth that it

has partial independence and it is composed by partisans, and fifth that it has partial

independence and is composed by technocrats. Each one of these combinations of composition

and level of independence produces a different time trajectory of money growth as a function of

the composition of government and the timing of elections: left wing governments want

expansion of monetary growth; partisan banks try to help a government of their party and hurt an

opposed government particularly at the time of elections, while technocrats behave the same way

regardless of who is in power; independent banks are not affected by approaching elections,

while non-independent are.

In order to test all her hypotheses Lohmann introduces a series of variables: economic

(monetary growth, GNP growth, inflation, exchange rates), institutional economic (Bretton

350

Woods, European Monetary System), electoral cycles, government composition, upper house

(Bundesrat) composition, central bank composition (who appointed the different members), and

chancellor popularity. Her results support the following two conclusions: “German monetary

policy is subject to electoral pressures. There is no evidence that partisan preferences are

influential via the power of appointment. The Bundesbank Council is stuffed with partially

independent technocrats whose independence decreases with the partisan support for the federal

government in the Bundesrat….. The auxiliary hypothesis that the Bundesrat veto point protects

the Budesbank’s independence is the only hypothesis that is consistent with the evidence

compiled in both the case study and the regression analysis” (Lohmann 1988: 440)

Lohmann’s findings are consistent with the theory presented in this book. However, as

she notes, the composition of the Bundesrat in Germany is correlated with government

popularity, so the institutional measures might in fact reflect the ability of government to control

monetary policy. This is a problem of co-linearity that cannot be resolved with a dataset covering

one country and forty-five years.

However, two more recent studies (Keefer and Stasavage (2000a and 2000b)) have much

more expanded datasets. Philip Keefer and David Stasavage (2000a: 17) develop an economic

model that leads to the following three predictions: “1. The presence of a legally independent

central bank should have a negative effect on inflation only in the presence of checks and

balances. 2. Political interference, such as replacement of central bank governors, is less likely

when checks and balances are present. 3. The presence of a legally independent central bank has

a more negative effect on inflation when different branches of government have divergent

preferences over inflation.” All three predictions are consistent with the theory developed in this