Hvg krause 47K. qxd

---

Veto Points, Policy Preferences, and 

Bureaucratic Autonomy in Democratic Systems

Thomas H. Hammond

In the development of democratic systems over the past two centuries, a

major thrust has been to remove policy-making power from the hands of

autocrats and their bureaucrats and place it in the hands of elected po-

litical leaders. Hence, it is an irony of twentieth-century governance that,

as the social and economic responsibilities placed on democratic govern-

ments have increased, their elected political leaders have responded by

delegating increasing amounts of policy-making authority back to un-

elected of

ficials—in particular, back to the bureaucrats.

While the expansion of the modern welfare state has occasionally stim-

ulated claims that modern bureaucrats have become our new rulers, it

would be a gross exaggeration to say that the bureaucrats in democracies

are generally able to operate independently from elected of

ficials. None-

theless, these bureaucrats often have at least some independent policy-

making power, whether explicitly delegated to them or not. Hence, it is

important to determine the conditions under which bureaucrats in

democracies have more independent policy-making power and when they

have less.

There is some evidence that bureaucrats—both within and across

democracies—do vary in the extent to which they are controlled by

elected of

ficials. Regarding evidence for differences within democracies,

for example, there is general agreement that the Federal Reserve System

in the United States is able to operate more independently from the pres-

ident and Congress as it makes monetary policy than are, say, the De-

partment of Agriculture and the Agency for International Development

73

---

Politics, Policy, and Organizations

as they make agricultural and international aid policies. Evidence for

these kinds of differences across democratic systems can be found in

studies such as Aberbach, Putnam, and Rockman 1981 and Weaver and

Rockman 1993.

These varying degrees of bureaucratic autonomy can have important

consequences for the policies that emerge. The reason is that the policy

preferences of the bureaucrats are not always representative of the policy

preferences of their elected overseers (see Aberbach and Rockman 2000

for the United States and Aberbach, Putnam, and Rockman 1981 for ad-

ditional democracies). To the extent that the bureaucrats have policy

preferences that differ from those of the elected of

ficials, and to the ex-

tent that the bureaucrats also have some independent policy-making ca-

pability, then to that extent their policy choices will differ from those of

the elected of

ficials.

There are two major reasons why the bureaucrats might develop an in-

dependent policy-making capability. One reason, which follows the tra-

dition established by Weber, is that the bureaucrats may know more than

the elected of

ficials about what needs to be done and how to do it. Even

if elected of

ficials do not explicitly delegate policy-making authority to

the bureaucrats, the bureaucrats’ greater information, theoretical under-

standing, and operating expertise may give them scope for independent

action.

The other reason why bureaucrats may develop a capability for inde-

pendent action is related to an age-old strategy for victory in both war-

fare and politics: “Divide and conquer!” If the elected of

ficials are di-

vided among themselves (i.e., if they have differing preferences over what

they want the bureaucrats to do), then the bureaucrats may be able to

conquer (in the sense of maintaining some scope for independent ac-

tion). But if the elected of

ficials are unified in what they want the bu-

reaucrats to do, the bureaucrats may have to do what they are told.

Most discussions of bureaucratic autonomy have focused on the au-

tonomy that stems from the asymmetries in information, understanding,

and expertise; less attention has been paid to the bureaucratic autonomy

that may result from divisions among the elected of

ficials. However, re-

cent work—see, for example, Hammond and Knott 1996, 1999, 2000,

which build on Hammond and Miller 1987—has begun to explore the

extent to which divisions among the elected of

ficials in the United States

can also allow bureaucrats some autonomy. This essay further examines

74

---

the extent to which divisions among elected of

ficials may generate au-

tonomy for bureaucrats, both within and across democracies.

There are two different kinds of divisions among politicians that might

affect bureaucratic autonomy. One involves the political veto points that

can reject proposals for policy change. For example, in analyzing why dif-

ferent kinds of political systems respond in different ways to social and

economic challenges, Weaver and Rockman (1993) discuss why two-party

unicameral parliamentary systems (the “Westminster” systems) might re-

spond more quickly than presidential systems (the separation of powers

systems, as in the United States). The basic argument is that because pres-

idential systems are characterized by multiple veto points (e.g., a House,

Senate, and president) whose members are able to block attempts at pol-

icy change, the systems’ responses to social and economic challenges are

often problematic. In contrast, the existence of only a single veto point—

the majority party—in Westminster systems enables these systems to re-

spond to challenges more readily.

However, Weaver and Rockman (1993) 

find that this institutional ex-

planation is insuf

ficient to account for the patterns of policy change that

they empirically observe. Instead, a second kind of division among

politicians also seems important, focusing on the extent of policy dis-

agreements among the elected of

ficials within and among the veto points.

When policy disagreements across institutions are modest, a system’s ca-

pacity for responding quickly to social and economic challenges may

seem to be greater than when policy disagreements among the institu-

tions are more substantial.

These insights have been formalized in Tsebelis 1995 and Tsebelis and

Money 1997, which advance two general arguments. First, holding con-

stant the extent of policy disagreements, an increase in the number of

veto points will not decrease policy stability and may increase it. The

logic here is straightforward: the more veto points there are the more

dif

ficult it is to gain approval for a policy change. Second, holding con-

stant the number of veto points, an increase in the extent of policy dis-

agreements among the actors will not decrease policy stability and may

increase it. Empirical research by Tsebelis (1999) and Bawn (1999) pro-

vides support for this general line of argument.

However, Hammond and Butler (2003) caution that both variables

must be considered in evaluating the extent of policy stability in any par-

ticular kind of system: just by itself, the number of veto points may not

Veto Points in Democratic Systems

75

---