Weingast B.R., Wittman D. The Oxford Handbook of Political Economy

Подождите немного. Документ загружается.

chapter 1

..............................................................

THE REACH OF

POLITICAL

ECONOMY

..............................................................

barry r. weingast

donald a. wittman

Over its long lifetime, the phrase “political economy” has had many diﬀerent mean-

ings. For Adam Smith, political economy was the science of managing a nation’s

resources so as to generate wealth. For Marx, it was how the ownership of the means

of production inﬂuenced historical processes. For much of the twentieth century, the

phrase political economy has had contradictory meanings. Sometimes it was viewed

as an area of study (the interrelationship between economics and politics) while at

other times it was viewed as a methodological approach. Even the methodological

approach was divided into two parts—the economic approach (often called public

choice) emphasizing individual rationality and the sociological approach where the

level of analysis tended to be institutional.

In this Handbook, we view political economy as a grand (if imperfect) synthe-

sis of these various strands. In our view, political economy is the methodology of

economics applied to the analysis of political behavior and institutions. As such, it

is not a single, uniﬁed approach, but a family of approaches. Because institutions

are no longer ignored, but instead are often the subject matter of the investigation,

this approach incorporates many of the issues of concern to political sociologists.

Because political behavior and institutions are themselves a subject of study, politics

also becomes the subject of political economy. All of this is tied together by a set

of methodologies, typically associated with economics, but now part and parcel of

4 introduction: the reach of political economy

political science itself.

The unit of analysis is typically the individual. The individual

is motivated to achieve goals (usually preference maximization but in evolutionary

games, maximization of surviving oﬀspring), the theory is based in mathematics

(often game theoretic), and the empirics either use sophisticated statistical techniques

or involve experiments where money is used as a motivating force in the experiment

(see Palfrey, this volume).

The purpose of this Introduction is to illustrate the intellectual excitement in

political economy by covering some important elements on the scholarly frontier.

As such, it neither provides an outline of the volume nor a summary of the major

topics and results. In this chapter, we discuss a set of approaches and issues that have

spawned interesting results and that are likely to spur considerable research in the

next decade.

We divide our essay into ﬁve sections. In Section 1, we discuss research on endoge-

nous institutions. The research agenda on institutions follows a natural progression.

The ﬁrst step is to determine how institutions aﬀect behavior. Indeed, this step seems

a necessary condition for a theory of endogenous institutions. Having built up a large

literature on the eﬀects of institutions, students of political economy have begun to

treat institutions as endogenous (thereby incorporating some of the subject matter

of sociology and anthropology). We focus our attention on legislative institutions

because this is where much of the work has been done. The success of institutional

analysis of legislatures is not surprising, as scholars have collected a large body of data

and evidence (both quantitative and qualitative) on legislatures. For example, votes

have been recorded with party aﬃliations and other attributes noted. These large data

sets allow hypotheses to be tested and theory to be reﬁned. Because the rules of the

US Congress are internal to Congress, voting procedures, the type of committees,

and committee assignments are all endogenous. So legislatures are fertile ground for

exploring institutional choice.

One of the technically most challenging but at the same time one of the most

exciting areas of research in political economy concerns the revelation and aggre-

gation of information, the subject of Section 2.Thisworkisexcitingbecausemanyof

the results contradict earlier beliefs based on decision-theoretic models and because

this research answers many puzzles. Here, our focus is on voters, particularly voters

who are uninformed in one way or another, but are nevertheless rational. Since this

research area is still in its infancy, we expect much more to be done in the ensuing

years.

Section 3 is devoted to evolutionary models of human and political behavior.

Political economy is now at the conﬂuence of two related paradigms: utility maxi-

mization and evolutionary ﬁtness. Both employ survival arguments in the context

of competitive forces—for example candidates need to win elections to survive. And

both employ the concept of equilibrium. These two concepts of survival and equi-

librium distinguish political economy from other approaches to political behavior.

However, these two approaches at times provide contradictory insights. As we will

¹ See Austen-Smith, this volume.

barry r. weingast & donald a. wittman 5

show, some kinds of irrational behavior may improve evolutionary ﬁtness. So at

the same time that political economy is pushing the envelope of hyper-rationality

(asillustratedinSection2), it is also trying to incorporate elements of emotions

and irrationality (Section 3). Furthermore, while political economy has tradition-

ally been based on self-regarding behavior, a considerable body of research in evo-

lutionary politics tries to explain other-regarding behavior, such as altruism and

vengeance.

Scientiﬁc knowledge depends to a great extent on the interplay between empirical

knowledge and theoretical development. Not surprisingly, our most comprehensive

knowledge is about the advanced industrial democracies in general and legislatures

in particular, where the great number of observations (of votes, party aﬃliation, etc.)

allow for an extensive testing of hypotheses and considerable reﬁnement of theory.

Our Handbook of Political Economy reﬂects this emphasis.

Nevertheless, over time, there has been a spread of knowledge from the core

areas of research. This spread has occurred for several reasons. First, the same be-

havioral relations that we observe within democracies may occur across political

systems once we account for the divergent institutional constraints on the actors.

For example, authoritarians may not face elections, but they too need political

support to remain in power (see Bueno de Mesquita, this volume). Second, more

information is being collected so that cross-country comparisons can now be done.

Finally, the political phenomena in non-democratic countries raise a host of ques-

tions typically ignored in democratic countries that demand answers: why is there

ethnic conﬂict? When is democracy a stable political system? What if any is the

relationship between democracy and capitalism? And why are so many nations

underdeveloped?

In Section 4, we consider the spread of political economy to new areas of research.

Here the empirical and theoretical answers are the least certain, but perhaps the most

interesting because of their novelty. We use, as our illustrative example, work on the

size and wealth of nations. A motivating reason for choosing the size of nations as our

prime example of the spread of political economy is that rational choice models have

often been (unfairly) accused of dealing with “epiphenomena” such as voting rather

² This disproportionate focus of political economy research has arisen for several reasons. First, the

political economy tools were ﬁrst developed studying democratic countries and are therefore more easily

adapted to other democratic countries than to non-democratic ones. Second, close observation and data

are more easily obtained in democratic countries so that theories applying to them have been honed the

most. Third, the institutional tools of political economy are more readily applied to the more highly

developed institutions of the advanced industrial democracies, in contrast to the less stable and less

institutionalized politics in the developing world.

³ Indeed, another deﬁning characteristic of the political economy approach is the use of large data

sets that enable econometric comparisons across a variety of countries, where the varieties are captured

by diﬀerent independent variables. For examples of cross-country comparisons, see Persson and

Tabellini (this volume) and Glaeser (this volume). The econometric approach is in stark contrast to the

older comparative politics literature, which compared two or three countries at a time.

⁴ For Handbook surveysoftheseﬁeldssee,respectively:Fearon(thisvolume),Przeworski(this

volume), Iverson (this volume), and both Acemoglu and Robinson (this volume) and Bates (this

volume).

6 introduction: the reach of political economy

than with “deeper and more substantive” issues. The size and wealth of nations clearly

passes the gravitas test.

1 Endogenous Institutions:The

Structure of Congress

.............................................................................

Institutions can be studied at three diﬀerent levels. First, the most basic and common

level takes institutions as given and studies their eﬀects. Second, the ﬁrst method

can be used as a form of comparative institutional analysis to study the implications

of diﬀerent forms of institutions. Third, the deepest level of institutional analysis

is to take the institutions themselves as endogenous; and to explain how and why

institutions are structured in particular ways, and why some types of institutions

survive but not others. The third approach is both the newest and the least explored

of the three approaches to institutions and is therefore likely to be a major frontier in

the coming years.

To illustrate the diﬀerences in the three approaches to institutions, we focus on

legislatures, where scholars have made signiﬁcant progress on institutional choice. We

begin our discussion with models that take legislative institutions as given and study

their eﬀects. In a relatively ‘institution-free’ legislature with majority rule voting in

one dimension, legislative choice will be the preference of the median legislator. By

adding institutional features to this simple spatial model of legislative choice, scholars

have studied the implications of a variety of institutional details. For example, several

scholars have studied the eﬀect of committee gatekeeping authority (Denzau and

MacKay 1983; Shepsle and Weingast 1981)orparty gatekeeping authority (Cox and

McCubbins 2005) on legislative choice. The idea is that committees (during the mid-

twentieth century) or parties (during the late twentieth century and early twenty-ﬁrst

century) held the power to keep issues within their jurisdiction from coming up for a

vote. In contrast to the median voter model, the gatekeeping models show that some

non-median status quos can be sustained: the gatekeeper will keep the gates closed

for any status quo that she prefers to the median’s ideal. This research agenda has

produced a wealth of knowledge about how legislative institutions aﬀect both legisla-

tive policy choices and policy decisions by the other branches. For example, Laver and

Shepsle (1996; see Laver, this volume) show how parliamentary institutions aﬀect pol-

icy choice; Krehbiel (1998, this volume) demonstrates the eﬀects of internal congres-

sional rules (notably, the ﬁlibuster) for policy-making, including how policy-making

changes with various types of electoral change; Ferejohn and Shipan (1990;seeHuber

and Shipan, this volume) show how potential threats of legislation aﬀect bureau-

cratic decision-making even without any legislation passing; and Marks (1988;see

McCubbins and Rodriguez, this volume) demonstrates the close relationship between

legislative preferences and judicial decisions that interpret the meaning of statutes.

barry r. weingast & donald a. wittman 7

The second type of institutional analysis utilizes the above methodology to make

comparative statements about diﬀerent institutions. For example, pivotal politics

models show the diﬀerences in behavior between a unicameral majoritarian system,

a bicameral system where each chamber uses majority rule, and a bicameral system

in which one chamber employs a ﬁlibuster rule allowing a minority of legislators to

prevent the passage of legislation. The ﬁrst institution always results in the median

legislature’s ideal policy. The second institution creates a gridlock range of possible

policy choices—the set of points between the ideal policies of the median in each

chamber. Any status quo policy in this set is an equilibrium in that there does not

exist a majority in each chamber that can overturn it. The third institution extends

the set of status quo points that cannot be overturned even further: the possibility

of a ﬁlibuster means that only policies commanding 60 per cent in one chamber and

a majority in the other can overturn a policy, so more status quo policies are stable.

For a further discussion of these issues see Krehbiel (this volume) and Cutrone and

McCarty (this volume).

Third, a much smaller set of papers studies the structure of the legislature itself and

treats its institutions as endogenous.

Four diﬀerent approaches have been used to

explain legislative structure: (1) legislator preferences, (2) committees as commitment

devices, (3) parties as transactions cost reducers, and (4) committees as information

providers.

We will now discuss each in turn.

1.1 Legislator Preferences

The simplest of the approaches bases legislative choice on legislator preferences

and relies on the “majoritarian postulate,” which holds that legislative policy and

procedural choices are made by majorities (Krehbiel 1991). In the context of

one-dimensional models of policy choice, the preference-based approach has the

following implications (Krehbiel 1993). First, policy choice corresponds to that of

the median legislator. Second, suppose that legislators join one of two parties, and,

further, that those to the right of the median largely join one party while those to the

left of the median largely join the other party.

Suppose that the status quo is to the right of the median and the proposed

legislation seeks to move policy left toward the median voter’s ideal. The “cutting

line” divides the set of voters into those favoring the status quo and those favoring

the proposal. In this context, the cutting line is that policy halfway between the

status quo and the proposed alternative (assuming that legislator utility functions

are symmetric). Since the status quo is to the right of the median, so too will be the

cutting line. The proposal makes every legislator to the right of the cutting line worse

oﬀ, so they vote against the policy; while every legislator to the left of the cutting line

is better oﬀ under the proposal and will vote for it.

⁵ We ingast 2002 surveysthismodeofanalysisindiﬀerent contexts. Shepsle (this volume) provides a

variantonthesethemes.

⁶ Laver (this volume) covers some of these issues; see also the summary in Shepsle and Weingast 1995.

8 introduction: the reach of political economy

Given the assumption of how legislators choose parties, nearly all legislators from

theleftpartyvotefortheproposal;whilemostofthoseoftherightpartyvoteagainst

it. Indeed, if the status quo is not too far to the right relative to the distribution of

legislature preferences, then most of the members of the right party will vote against

the change. In other words, voting on this legislature will exhibit polarization by party

even though the party exerts no pressure on its members to vote one way or another.

Alessonofthismodelisthatpolarizedpartyvotingcanemergeasthecombined

result of legislative preferences and sorting into parties without being a function of

any legislative institutions that advantage parties or that constrain member behavior.

Although this approach rationalizes only minimalist legislative institutions, it

provides an important baseline from which to judge other models. While most

approaches rely on legislator preferences to some degree, we term this approach a

preference-based approach because it relies solely on legislative preferences and the

median voter model to explain political phenomena, such as polarized party voting.

1.2 Committees as Commitment Devices

The second approach is exempliﬁed by Weingast and Marshall’s (1988) “Industrial

organization of Congress.” This approach built on previous theoretical and empirical

work. Going back to Buchanan and Tullock (1962), many models of legislative choice

emphasized logrolling and vote-trading. By logrolling and trading votes, members

and the districts they represented were better oﬀ. Logrolling can thus be seen as a

legislative institution parallel to market institutions in the economic sphere.

Empirically, the substantive literature on Congress emphasized the central impor-

tance of committees, which were seen to dominate the policy-making process. That

literature emphasized committee specialization and self-selection onto committees

by members most interested in the committee jurisdiction (Fenno 1966, 1973; Shepsle

1978). Clearly, this form of committee organization suited members’ electoral goals

(Mayhew 1974). So committee organization was seen as reﬂecting the preferences of

the legislators and their constituents. In this view, the key to understanding legislative

organization was legislative exchange.

Weingast and Marshall sought an explanation of congressional organization that

accounted for the fundamental features then found in the substantive literature. They

based their approach on two observations: ﬁrst, the legislature faced many diﬀerent

issues that cannot be combined into a single dimension: agriculture is not commen-

surate with civil rights, banking, or defense. Second, vote-trading had signiﬁcant en-

forcement problems as a means of legislative exchange. For example, suppose that one

group of legislators seeks to build dams and bridges, another group seeks regulatory

control of some market, and that neither group alone comprises a majority. The two

groups could, per logrolling, agree to support one another’s legislation. But this raises

a problem: once their dams and bridges are built, what stops those receiving them

from joining those locked out of the original trade to renege on the deal by passing

barry r. weingast & donald a. wittman 9

new legislation ending the regulation? Because of the possibility of reneging, some

logrolls will fail ex ante as legislators fear their deals will ultimately fail.

Enforcement problems imply that direct exchange of votes is not likely to provide

a durable means of legislative exchange. Instead, Weingast and Marshall argued that

legislators were likely to institutionalize their exchanges in the form of a legislative

committee system (LCS) that granted legislators greater powers over policies within

the committee’s jurisdiction. They showed that, in the context of a mechanism to

grant rights to committee seats in combination with self-selection onto committees,

the LCS made members with diﬀerent preferences better oﬀ.

Consider the problem of reneging noted above. Self-selection onto committees

with gatekeeping power prevents this type of reneging. Suppose the group favoring

dams and bridges seek to renege on their original deal and introduce legislation to

undo the regulation. This legislation now goes to the committee with jurisdiction.

Populated by those who favor maintaining the regulation, committee members pre-

vent the legislation from coming before the legislature. This preserves both the status

quo and the original legislative exchange.

This approach also addresses an important question raised by the majoritarian

postulate. This postulate questions why a majority would ever vote to reduce or

restrict its own powers in the future. In the context of a single dimension of legislative

choice, it is hard to understand why the median (and hence a majority) would vote to

restrict itself. In the context of multiple dimensions, however, no median exists. The

exchange postulate underlying the LCS provides an answer to the question: a majority

votes to restrict itself on a series of diﬀerent policy issues simultaneously. Although

this restricts the majority’s actions on each dimension, if each member is assigned to

a committee of higher value than the average, this exchange makes each better oﬀ (see

also Calvert 1995).

In this model, committee organization solves the problem of legislative exchange.

Given pervasive enforcement problems of direct exchange of votes, legislators instead

choose to organize the legislature in such a way as to institutionalize a pattern of

exchange that furthers the goals of all.

1.3 Legislative Parties as Solutions to

Collective Dilemmas

The third approach uses legislative parties to explain legislative organization and

behavior. In this view, parties are more than just a collection of people choosing

the same party label. Cox and McCubbins (1993), for example, argue that legis-

lators face a series of collective action problems that political parties can resolve.

For example individual legislators have trouble passing their own legislation; and

without coordinating, legislator activity fails to add up to enough to help each get

⁷ A second enforcement problem is that exchange of votes over time creates additional opportunities

for reneging, especially as bills evolve.

10 introduction: the reach of political economy

re-elected. In particular, all legislators face a common-pool problem in which they

have incentives to shift costs onto each other. Parties overcome these problems by

enforced coordination.

In the face of various coordination and related problems, Cox and McCubbins

argue that members have an incentive to use parties to coordinate the behavior of

their members for several ends: to produce legislation more attractive to their mem-

bers; to develop a national reputation or brand name; and, in combination, to use

these tools to help re-elect their members (see also Wittman 1989, 1995).Committees

in this view are a tool of the majority party used to further party goals; namely, to

propose legislation beneﬁting party members and to prevent legislation that would

make party majorities worse oﬀ. The majority party’s delegation to each committee,

rather than being composed of those most interested in the policy as in the Weingast

and Marshall approach, were representative of the party. This particularly holds for

gatekeeping committees, such as the budget committee where the members do not

self-select. Tests of the representatives of committees tend to support the party view

(see, e.g. Cox and McCubbins 1993). Also consistent with this view was the striking

partisan aspect of congressional voting, particularly since 1980.

However, this party-centric approach has not been without its critics. Krehbiel

(1993) presented a major challenge to this perspective by asking, ‘where’s the party?’

Relying on the preference-based approach noted above, he showed that many of

the ﬁndings of the party-centric perspective were consistent with the majoritarian

perspective. We have already noted how polarized party voting, rather than being a

product of the party organization of the legislature, can result from simple preferences

in combination with legislator sorting into parties. Another aspect is the representa-

tiveness of committees. As noted, the party perspective emphasizes that each party’s

delegation to a committee is representative of the party; but if both parties do this,

then the overall committee will be representative of the chamber, also consistent with

the majoritarian perspective.

The debate about parties has spurned a remarkable empirical literature. See for

example, Cox and McCubbins (2005), Krehbiel (this volume), and Groseclose and

Snyder (2001). We do not have time in our introduction to cover this literature, but

we do want to emphasize that the research in endogenous legislative institutions is

empirical, as well.

1.4 Information Explanations for Structure

The ﬁnal approach to legislative organization, associated with Gilligan and Krehbiel

(1989)andKrehbiel(1991), emphasizes that legislators are uncertain about the im-

pact of their choices on actual outcomes. Legislators therefore have an incentive to

organize the legislature to reﬂect the task of gaining expertise and information that

⁸ See also Cox (this volume) who shows how legislative parties arise endogenously as a means to

resolve the problem of potential overuse of plenary time.

barry r. weingast & donald a. wittman 11

reduces this uncertainty. In this world, committees are bodies of legislative experts in

the policies of their jurisdiction. Committee expertise allows committee members to

reduce the uncertainty between legislation and actual outcomes.

This perspective has signiﬁcant implications for legislative organization, including

the choice of rules governing consideration of legislation on the ﬂoor. For example,

because expertise requires costly investment, legislators will undertake this costly

investment only if the system somehow compensates them for this. Krehbiel argues

that restrictive rules that bias legislative choice in favor of committees are the answer.

Although restrictive rules prevent legislators from choosing policy associated with

the median voter ex post,legislatorsarebetteroﬀ ex ante because committee expertise

allows committees to reduce the uncertainty associated with the diﬀerence between

legislation and policy outcomes.

1.5 Concluding Thoughts

The debate about legislative institutions has been lively, and no consensus has yet

emerged on the determinants of legislative organization. We cannot yet say whether

one perspective will ultimately triumph (as Gilligan and Krehbiel 1995 suggest) or

whether a synthesis of perspectives is likely to emerge (as Shepsle and Weingast 1995

suggest).

From a broader perspective, the study of legislative institutions provides a template

for how research on institutions is likely to proceed in the future. The ﬁrst stage

is to see how a particular institution aﬀects behavior; next, similar but somewhat

diﬀerent institutions are compared; then in the ﬁnal stage, institutions are treated as

being endogenous. If the history of research on endogenous legislative institutions

is any guide, there will be disagreement on which institutions are endogenous to

other institutions. These controversies, in turn, help shape our understanding of

institutions and provide a deeper understanding of organizations.

2 Revelation and Aggregation of

Information:Voting

.............................................................................

In this section, we consider the revelation and aggregation of information. This is

a game-theoretic, as opposed to a decision-theoretic, approach to information. An

exciting aspect of this research is that it often turns the standard theoretic wisdom on

its head. We illustrate by looking at voting behavior.

Traditional democratic theory argues that, for democracy to work, voters should

inform themselves about the candidates and the issues. Moreover, voters should

be unbiased and rely on unbiased sources of information. Practice in all working

⁹ Ansolabehere (this volume) reviews the broad topic of voting behavior.