‘ 'K‘
i: d .,
1.5.. nabflchnwwflshd
u I g xle‘ .‘l

 

 

 

 

 

 

I!
{5.15 ’1

41 lav
.l \

 

 

LIBRARY
Michigan State
University

 

This is to certify that the
dissertation entitled

AMBITION THEORY AND LEGISLATIVE ORGANIZATION

presented by

GREGORY ROBINSON

has been accepted towards fulfillment
of the requirements for the

Doctoral degree in Political Science

 

 

W0, m

Major Professor’s Signature
91-10—- 0 7

Date

 

MSU is an affinnative-action, equal—opportunity employer

PLACE IN RETURN Box to remove this checkout from your record.
TO AVOID FINES return on or before date due.
MAY BE RECALLED with earlier due date if requested.

DATE DUE , DATE DUE DATE DUE

 

 

 

AMBITION THEORY AND LEGISLATIVE ORGANIZATION
By

Gregory Robinson

A DISSERTATION

Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

DOCTOR OF PHILOSOPHY
Department of Political Science

2007

ABSTRACT
AMBITION THEORY AND LEGISLATIVE ORGANIZATION
By
Gregory Robinson

In this dissertation, I present a theory that posits political ambition as an important
factor shaping legislative and electoral behavior, as well as the endogeneity of the
internal structures of legislatures. A crucial distinction is made between Schlesinger’s
(1966) ambition theory and the reelection imperative, which is a special case of am—
bition theory.

I test hypotheses derived from this theory in three essays. The first deals with
the relationship between committee service and progressive ambition in the United
States House of Representatives. Based largely on the activities argued by Mayhew
(1974) to facilitate the pursuit of higher office, I present the concept of ‘the progressive
ambition comrnittee’ and argue that the House Judiciary Committee is an existing
example of this concept.

The second essay tests a prediction of partisan theories of legislative organization,
that progressively ambitious House members will adjust their roll call behavior to
accommodate the statewide constituency they hope to win, but that this accommo-
dation does not extend to votes on special rules, which are the principal tool by which
the majority party controls the agenda. The theorized motivation for the loyalty of
progressively ambitious members on these procedural votes is the protection of party
reputation, a collective good from which a Senate candidate can still benefit.

The third essay explores the implications of ambition theory for our understand-
ing of rules changes in legislative chambers. Focusing on how term limits structure
the career possibilities of state legislators, I examine how the replacement of static

ambition with other career concerns changes the incentives associated with institu-

tional maintenance. Specifically, I use term limits as a natural experiment to test a
hypothesis about the maintenance of committee property rights in state legislatures.

The contribution of the dissertation is to bring together insights from the largely
distinct literatures on legislative organization and strategic candidates by showing
how legislators use different aspects of the institution to pursue a wider set of goals

than previous scholars of legislative organization have allowed.

Copyright by
Gregory Robinson
2007

For those who crossed oceans to bring me here.

ACKNOWLEDGEMENTS

An education is a collective enterprise. The student makes advances because of
the help of some, and despite the hindrance of others. In the end, the student is in
debt to both. I

I could write at length about all the people, in either category, to whom I owe
debts. Instead, I will acknowledge those who have been of special help in writing
this dissertation, and in helping me get to the point where I could begin writing it.

Without the political science, philosophy, and anthropology faculty at Grand
Valley State University, I would not have become interested in graduate school, and
may very well not have finished my undergraduate education. Specifically, I must
thank Erika King, Roger Moiles, Deborah Orth, and Mark Richards in Political
Science, as well as Mark Moes and Leo Zaibert in Philosophy for encouraging my
interest in the study of politics and of political thought. Forrest Armstrong, who
passed away not long before I left for graduate school, was especially encouraging.

For being of such great help during that first, hectic year of graduate school
(along with much other help in later years), I would also thank Burt Monroe and
Dave Rohde. As supervisors for a journeyman research assistant, Eric Chang and
Paul Abramson were great bosses who taught me a lot about how to do political
science.

It‘would take too long to detail the enormous help my graduate student colleagues
have been. Jamie Carson, Mike Crespin, and Kirk Randazzo have been friends, col-
leagues, and mentors. Gia Barboza, Jeremy Duff, Suzanne Gold, and Ben Lundgren,
all fellow survivors of my graduate cohort, are comrades with whom I will always
share a special bond. Sasha Miller, Carl Snook, and Fred Wood, though they are
colleagues who joined us later, feel no less our comrades for it. I

As academics, we cannot thank enough the staff who keep the place running.

Crucial paperwork gets where it needs to go (as do our paychecks!) and we don’t

often take the opportunity to let them know how much we appreciate all that they do
for us. Thanks, then, to Karen Battin, Rhonda Burns, Kelly Fenn, Shirley Gordon,
and Char Rogers.

I also have to thank the organizers, the faculty, and the participants at EITM V
in Ann Arbor. It was a wonderful summer of friendship and scholarship (though I
admit that some of it was way over my head!) that I will treasure forever.

Adriana Cole and Jeff Holycross were of great help in collecting and compiling
the state legislatures data I used in the dissertation. Graduate students rarely get
the benefit of research assistance themselves, so I am grateful to have had their help,
and for their work to have been so excellent with almost no direction from me. They
did a wonderful job!

Stewart French deserves special thanks, as friend, comrade, and collaborator. We
already have more coauthored projects planned than we could finish in a lifetime,
but lets hope that means we will finish at least some of them!

I cannot thank the members of my dissertation committee enough. Tom Ham-
mond gave me tough, extensive, but above all incredible useful critiques of my work,
and I am extremely grateful for the time and effort he devoted to helping make my
dissertation as good as possible.

My discussions with Carlos Pereira not only made the current form of the dis-
sertation better, but have left me with years of work to do on a research agenda
that extends what I’ve already done. At least some of that work will hopefully be in
collaboration with Carlos, which is an exciting prospect.

Rick DeShon was a stupendous outside reader and was, again, extraordinarily
generous with his time and genuinely interested in work that was not in his field.

Nate Monroe is as great a dissertation chair and mentor as I could have asked
for. I can honestly say that I would not have finished graduate school if he had not

arrived at MSU when he did. My time working with him taught me what science is,

vii

and how we should apply that understanding to the study of politics. Much of the
credit for my Ph.D. and for my job placement goes to Nate. While I can list these
things and many more, I cannot really put in words What his friendship and support
have meant to me.

My father and brother were my family when I was growing up. Looking back on it
objectively, I might be tempted to say that our lives were pretty tough at times, but it
never really felt that way. And it still doesn’t. My dad has been quietly encouraging
these five years, though I think he would have been glad if grad school had allowed
me to visit more often. A reward for both of us as I finish this dissertation is that
I’ll get to spend the summer with him before I start my new job.

I had the great fortune this past year to have my brother as a roommate, which
has been a great experience. I hope I was as much a help to him in pursuing his
career in education as he was in keeping me sane through the process of being on
the academic job market and trying to finish a dissertation. Though I don’t get to
see them as often as I’d like, I cannot forget my mother and my sister who are more
excited than anyone else at having a ‘doctor’ in the family.

There are too many more people to thank, both living and dead, but most of

them know who they are. To sum up,

Let’s drink to those that wish us well. All the rest can go to hell.

viii

Contents

List of Tables xi
List of Figures xii
1 Introduction 1
1.1 My Theoretical Perspective ....................... 2
1.1.1 Neoclassical Economics, Neoinstitutional Economics, and Ra-

tional Choice Institutionalism .................. 3
1.1.2 The Central Argument of Rational Choice Institutionalism . . 6
1.1.3 What Rational Choice Institutionalism Examines ....... 7
1.2 A Unified Theory of Legislatures? .................... 10
1.3 Schlesinger’s Ambition Theory ...................... 13
1.4 Ambition Theory and the Study of Legislatures ............ 15
1.5 General Assumptions ........................... 17
1.6 Testing Theories of Legislatures ..................... 21
1.7 Competing Theories: What Ambition Contributes ........... 27
1.8 Organization ............................... 30

2 Progressive Ambition and Committee Membership in the House of
Representatives 31
2.1 Introduction ................................ 31

2.2 A Case Study in the Use of Committee Membership to Pursue Higher

Office .................................... 32
2.3 Assumptions & Hypotheses ....................... 38
2.4 Analysis .................................. 41
2.5 Substantiating Assumptions: Is the Judiciary Committee Unique? . . 47
2.6 Conclusion ................................. 50

3 Progressive Ambition and Partisan Theories of Legislative Organi-
zation 55
3.1 Introduction ................................ 55
3.2 The Danger of Sophisticated Legislating? ................ 57
3.3 Previous Work .............................. 59
3.4 Assumptions & Hypotheses ....................... 61
3.5 Analysis .................................. 66

ix

3.6 Conclusion .................................

4 Toward an Ambition Theory of Legislative Organization
4. 1 Introduction ................................
4.2 An Ambition Theory of Institutional Maintenance ...........
4.3 Hypotheses & Data Analysis .......................
4.3.1 Individual Committee Seat Property Rights ..........
4.3.2 Committee Jurisdictional Property Rights ...........
4.4 Discussion .................................

5 Conclusion
5.1 Summary .................................
5.2 What I Have Learned ...........................

Bibliography

70

73
73
76
81
83
87
91

93
93
95

99

List Of Tables

2.1

2.2

2.3

4.1

4.2

Committee Membership & Higher Office Ambition ...........

Average Number of Committee Hearings per Congress, 80th—108th
Congress ......... ’ ...........................

Average Number of Bills Reported from Committee per Congress,
80th—105th Congress .............................

Preference Change in the House of Representatives ...........
Committee Creation in the Michigan Legislature ............

Creation of New Committees in Four State Legislatures ........

46
48
50
69
86

90

List of Figures

3.1 Partisan Theories & Change in Constituency Preference ........ 63

Chapter 1

Introduction

Ambition is an ambiguous concept in the study of politics. Observers of politics
have used the term ambition, or related terms, for millennia. Translators of Aristotle
have often rendered one of his ten virtues, philotimz'a, or love of honor, as ambition in
English (see Thomson 1953). This usage, which implies a desire to be well thought of
by one’s fellow citizens, is somewhat related to but probably still far from the sense
in which most contemporary scholars of politics use the term ambition. Merriam-
Webster’s Dictionary defines the word ambition variously as “an ardent desire for
rank, fame, or power” or the “desire to achieve a particular end.” Taking these two
related definitions together gives us something closer to the sense in which James
Madison used the term ambition in Federalist 51.

Madison is often quoted in this regard, when he writes, “Ambition must be made
to counteract ambition. The interest of the man must be connected with the consti-
tutional rights of the place.” This is the justification, according to Madison, for the
structure of government embodied in the Constitution. Separation of powers and
checks and balances, terms taught in American civics classes to capture the pecu-
liarities of the governing structure of the United States, are the embodiments of the
Founders’ skepticism of the motives of ambitious politicians.

For Madison, though, ambition in this sense goes beyond solely a political phe-

nomenon. It is better understood as simply one manifestation of a universal human

trait. He writes,

It may be a reflection on human nature that such devices should be
necessary to control the abuses of government...T his policy of supplying,
by opposite and rival interests, the defect of better motives, might be
traced through the whole system of human affairs, private as well as

public.

For the contemporary scholar of politics, ambition is often the answer to the
question of why individual citizens become politicians. Again, this is a more general
sense of the term than I will use throughout this dissertation, but it begins to come
close to the heart of the matter, in that it relates ambition to the political career.
The winning and holding of political office have become foundational in our study of
parties, legislatures, interest group activity, and beyond. And this points us to the
methodological individualism of the contemporary study of politicians and political
elites.

It now seems strange to think that there was a time when scholars did not view
political institutions and the behavior within them as determined by the goal—oriented
behavior of individuals, but such a point of view is an innovation that is, at best, a

little over 40 years old in political science.

1.1 My Theoretical Perspective

I consider my own theoretical perspective to be grounded in methodological individ-
ualism, with a particular focus not just on how goal-oriented behavior is conditioned
by institutional rules and structures, but on how goal-oriented behavior is in part
aimed at shaping those rules and structures. This perspective can be called ratio-
nal choice institutionalism. General arguments from a rational choice perspective

ha‘Ve proved controversial at certain points in the development of political science,

2

largely on the basis of assumptions of rationality and self-interestedness (Simon 1947,
1957, 1958; Green and Shapiro 1994; Friedman 1996). Controversies in economics
have evolved along different lines, and one in particular has long fascinated me is
the question of whether, as Field (1981) put it in his critique of the North/ Thomas
model of the economic development of Western Europe from feudalism to capitalism,

whether “rules antedate the market”.

1.1.1 Neoclassical Economics, Neoinstitutional Economics,

and Rational Choice Institutionalism
North (1981) defines an institution as “a set of rules, compliance procedures, and
moral and ethical behavioral norms designed to constrain the behavior of individuals
in the interests of maximizing the wealth or utility of principals” (201-2).1 Inherent
in North’s thinking on institutions is the endogeneity of institutions. Institutions
emerge, change, and disappear as a result of the actions and interactions of the
self—interested actors who operate within them. While North and Thomas (1973) is
among the grandest in scale, attempting to model the whole historical evolution of
economies, this and other similar works fall within a long line of economic theory
that has attempted to describe the emergence of institutions as both cause and con-
sequence of the behavior of self-interested actors (see especially Coase 1937; Alchian
and Demsetz 1972). This is the point of view often referred to as the new institu-
tionalism, or neoinstitutionalism.

So is this view of institutions consistent with neoclassical economics? It certamly
bears a similarity, relying on the assumption of actors as rational maximizers of sub-
jective utility. As Bates (2000) points out, the principal contribution of neoclassical
economics to the study of institutions has been to shift the focus from unitary actors

(often of the analyst’s invention) to the analysis of individual actors in the tradition

 

1An alternative definition, less general than North's, is found in Shepsle and Bonchek (1997,
3.11): “[I]t is appropriate to think of an institution as a governance structure, permitting individual
ambition and organizational purpose to be blended.” [emphasis added]

of the methodological individualism of Hayek (1973) and Popper (1944, 1945, 1957).2
But in a critique of North and Thomas (1973) in particular, and of the new institu-
tionalism in general, Field (1981) suggests that the framework presented by North
and others is not consistent with neoclassical economics. He writes, “Neoclassical
economics...presupposes not three but four categories of exogenous variables: tastes,
technologies, endowments, and rules [emphasis in original] (195).

So for Field ( 1981), the attempt to endogenize institutions is to leave neoclassical
economics behind, and is a fruitless enterprise besides. Field (1981) captures the
North and Thomas (1973) model, and much of the new institutional enterprise, in
the following: “[A]s the underlying parameters of an economic system change, rules,
originally efficient, become inefficient. Rule change results: These changes can be
said to be caused by changes in the underlying parameters of the system.” (183-4).

He further unpacks the theory, writing:

Changes in the prevailing rules are...to be understood as resulting from
the aggregation of the decisions of economic agents maximizing their
short-term individual interests in response to changes in the underlying
parameters of the economic system, as reflected in the changing explicit

or implicit prices they face (Field 1981, 185).

And while Field makes no reference the new institutionalism practiced within the
discipline of political science, he provides a summation of his critique with particular
relevance to the discipline of political science and its consideration of institutions of

various sorts:

In explanatory exercises of the sort North and Thomas and other at-

tempt, one cannot avoid at certain stages the appeal to prevailing rules,

 

2This point will be reappear below, when the distinction between new and old institutionalism
is discussed. It is also an important point of departure from the dominant view in political science
up through even the 19608, drawing particularly from the Nationalist sociological perspective
pioneered by Parsons (1937).

customs, or social structures as something independent of and transcend-
ing the individual interests of those agents whose behavior they regulate.

[emphasis added] (188).

This is directly relevant to questions dealing with the relationship between and rel-
ative importance of culture or norms of behavior and formal political institutions in
politics (cf. Matthews 1959, 1960). Field’s critique of the new institutionalist line
of thought within economics may have been apt, as his appeal to the necessary ex-
ogenously given factors of Walras’s (1954) general equilibrium model seems to make
clear.3 If Walrasian orthodoxy is a necessary condition to be included under the
neoclassical rubric, then North, Thomas, and those who make similar arguments are
in danger of being ejected from the club. And if such is the case, and the devel-
opment of rational choice institutionalism in political science finds its source in the
new institutional economics, then rational choice institutionalism is born of heresy.

While heavily influenced by neoclassical economics and the new institutionalism
of economics, rational choice institutionalism is in one sense very different from both.
Citing Schumpeter’s (1961) statement that all economic theory must be grounded in
a non-economic base, Field (1981) writes, “If one accepts this conclusion, then one
also accepts that rules must logically antedate the market” (193). It is very much
more difficult to argue that rules logically antedate politics.4 While economics is a
particular sphere in which the peculiarities of human nature play themselves out,
politics is human nature, from a certain point of view. To stretch a paraphrase of an
already over-used quote by Clausewitz (1873), rules are the continuation of politics

by other means. And so far from being a. colony of either orthodox or apostate

 

3Field argues that at crucial points in their argument, North and Thomas (1972) fall back on
axogenously given rules (or on a ‘conservative principle’ of institutional inertia) when faced with
inconvenient evidence, thus apparently demonstrating the neoclassical justification for taking rules
as given. North’s later work on transaction costs seems to have been driven in large part by this
criticism.

4Though here, it may be possible to place Nature’s ‘rule’ of the advantage of the strong over the
Weak, attributed as a theory of justice in Plato’s The Republic to Thrasymachus, as an ‘institution’
prior to politics.

neoclassicists, rational choice institutionalism actually seems to have the best chance

at carrying the enterprise of endogenizing institutions to its fruition.

1.1.2 The Central Argument of Rational Choice Institution-
alism (and How It Differs from Traditional Institution-
alism)

According to Riker (1990), rational choice institutionalism “relates structure to par—

ticipants’ incentives” (178). For Riker, old institutionalism is associated with in-

terpretation, while rational choice institutionalism is associated with generalization.

Rather than identifying institutional forms and delineating their consequences for

politics, a deeper understanding of the determinants of institutional choice and the

ways in which institutions serve the interests of participants is sought.

The old institutionalism often presents ’tangible’ institutional structures, like the
Congress or the Court or the Bureau, as unitary actors, greater than the sum of
their parts, their rules constituting a moral order that is the fundamental factor in
reinforcing the stability of the structure itself (Selznick 1996). Such a view cannot,
however,-make any real contribution to the explanation of institutional change.

Me to the legacy of neoinstitutional economics, rational choice institutional-
ism takes institutions to be the structures within which individuals pursue utility
maximization. And following the economic new institutionalism of North (1981,
1990a) and Dixit (1996), rational choice institutionalism views the stability of insti-
tutional structures as resulting from the transaction costs associated with institu-
tional change, and the most ambitious works have attempted to analyze institutional
structure as, on some level, endogenous to the operation of the political system. It
is important here to avoid confusion as to the use of the terms endogenous and ex-
ogenous. Gould and Baker (2002), for example, use exogenous and endogenous as
empirical categorizations of institutions, with exogenous denoting institutions that

are constitutionally established, and endogenous denoting institutional structures

that can be unilaterally changed by a single branch of government. Analytically,
however, all institutions are potentially endogenous. Taking the United States in
1787, Finland in 1991, Japan in 1996, and New Zealand in 1996 as just a few empir-
ical examples, and the model proposed by Aghion, Alesina, and rIirebbi (2004) as an
illustrative hypothetical, even the most fundamental aspects of the institutional struc-
ture that makes up a political system can be endogenous, analytically speaking. That
an institution is fundamental to the political structure does not mean that it must
be taken as exogenously given. It can rather be argued (and should at minimum
be considered) that the transaction costs associated with changing a fundamental
institution are almost always prohibitively high.

Among the best applications of rational choice institutionalism are analyses of
the United States Congress. An illustrative example is Rohde (1991), which depicts
the Democratic Party reforms of the early—19705 as the aggregation of the decisions
of political actors who became dissatisfied with the distortion of outcomes that the
institutional structure was creating, and who Were Willing to bear the transaction
costs associated with adjusting the institution in order to more efficiently pursue their
preferred outcomes (North and Thomas 1973; North 1990a; Dixit 1996). Again, as
Gould and Baker (2002) point out, not all institutions are so easily changed as was
the Democratic caucus in the House of Representatives. But by keeping transaction
cost politics firmly in mind when analyzing institutions, we can thoroughly relate
structure to the objectives of participants as both cause and consequence (Riker

1990).

1.1.3 What Rational Choice Institutionalism Examines, and
Some Proposals for What Else it Should Take a Look
At

Rational choice scholars analyze the institutional strength of legislative parties (Cox
1987; Rohde 1991; Aldrich and Rohde 2000), the formation of coalitions and cabi-

nets (Tsebelis 1991; Laver and Shepsle 1996), voting (Cox 1997), and so on. Much
of the institutional work in political science that has its roots in neoinstitutional
economics is still limited to dealing with institutional structures as causal factors in
the behavior of political actors (Tsebelis 1991; Cox 1997). As Cox (1997) states, “In
a world without transaction costs, democratic politics is inherently unstable” (269).
For North (1990b), transaction costs are the result of (or maybe even consist of, fun-
damentally) uncertainty. He thus provides an answer to questions raised by Coase
(1960), and proposes that institutions are created and adjusted to solve the problem
of transaction costs. So without transaction costs, all exchange, and all politics, is
unstable. The approach of Walras (1954) cannot explain the real-world operation of
either economics (Coase 1960) or politics (Arrow 1951; Shepsle 1979; Shepsle and
Weingast 1981).5 The oversight in much of the institutional political science litera—
ture is in the failure to recognize that institutions are chosen, at some point in time
by some set of actors, for reasons that ought to be explicable and eaplicated.

In the future research section of the concluSion to Making Votes Count, Gary
Cox refers to “the institutional engineering question” (277). For Cox, coordination
problems in politics are inevitable, and by picking a particular institutional formula,
institutional engineers can do much to determine when in the political process coor-
dination problems will arise, and what form they will take. Every institution favors
the interests of some subset of actors, and institutions take on their particular forms
precisely because of the ability of that subset to influence the engineering of the
institution. Giving short shrift to this idea leaves half of the institutional question
unanswered. So are Cox (1997), Tsebelis (1991), and others wrong in not focusing on
the institutional engineering question? No. An analytic and empirical understanding

of the operation of political institutions is invaluable, and can point out the potential

 

5The suggestion in North (1990b) is that the reason the neoclassical general equilibrium model
takes rules as exogenously given is the belief that, while they empirically do exist, they don’t really
matter.

for unintended consequences, which are an important part of the uncertainty upon
which transaction cost theories are based. But despite the importance of analyzing
the consequences of institutions, rational choice institutionalism in political science
needs to do more to endogenize institutions as consequences themselves.

Institutions are fundamentally important as explanatory factors in political sci-
ence. But just as important is the need for institutions themselves to be explained.6
Much has been done in the new institutionalism of economics to pave the way, and
political scientists of all fields should follow. Institutions are not exogenously given,
no matter what our theories must assume for the sake of convenience. Recognizing
this does not invalidate the voluminous research that has taken institutional struc-
tures and rules as given, but this recognition can help reinforce the more general
lesson that we must always be aware of our assumptions.

As I will go on to detail, there are many assumptions that could be reexamined.
Particularly, I will argue that our theoretical assumptions about politicians’ goals
should be loosened to account for some of the variation we observe in the world,
and that this broader set of assumptions can allow us to answer some important
questions. But at the same time, this question of the endogeneity of institutions is

implicit in the literatures with which I interact and in the arguments I will present.7

 

6See Aldrich, Jillson, and Wilson (2002), Cam and Shepsle (1989), Humes 1989, and Katz and
Sala (1996) for a few examples of this sort of analysis.

7Though how this endogeneity interacts with my proposed broader understanding of politicians’
goals will mostly be saved for Chapter 5. It is largely in service of the argument in Chapter 5 that the
preceding discussion of the new institutional economics is relevant, as the response of economic rules
and institutions to exogenous changes in the enviromnent is a useful analogy for the effect of term
limits on legislative structure, imposing, as they do, a new and radically difl’erent distribution of
ambitions on members of the term limited legislature. If ambition, and the distribution of individual
legislators into the different categories of ambition, is important for how particular legislatures are
organized, then a. shock to that distribution should have predictable organizational consequences.
While not directly analogous to the sorts of exogenous shocks that drive the North—Thomas model
(natural disasters, disease, climate, natural resources, conquest and invasion) I think the similarity
in the mechanisms by which exogenous shocks reveal the endogeneity of rules and institutions is
illustrative.

1.2 A Unified Theory of Legislatures?

Continuing along this line of thinking about assumptions, it is important to identify
and clarify the theoretical foundation of the study of legislatures. We start with
members of a legislature who have goals. The goals conceivably can be anything,
from career considerations to the accumulation of personal wealth to Aristotle’s ‘love
of honor’. The way scholars specify their assumptions about these goals will mean
much for the theories they build and for the conclusions they reach. Indeed, this
may be the biggest challenge for building a theory of legislative organization that
is generalizable across a wide set of political and social contexts. Building a theory
based on the assumption that legislators are single-minded seekers of reelection and
applying it in a context where, in reality, the widely-shared goal of legislators is the
extraction of rents (as political economists use the term-see Persson and Tabellini
2000) could well lead to erroneous conclusions. I will revisit the question of the
assumptions we make about legislators goals throughout this dissertation, since such
is the heart of what I call the ambition theory of legislative organization, but I will
leave the subject for now and proceed to the next critical aspect of the theoretical
foundation of legislative studies.

Legislators bring their goals to the arena in which they debate and disagree and
make policy, but there is a curious thing about the legislative arena. In nearly
every system, there are crucial aspects of the decisionmaking apparatus within the
legislature that can be established, eliminated, and otherwise changed by members of
the legislature. This is closely related to the argument put forward by Diermeier and
Krehbiel (2003), in which they distinguish between “institutional theories,” in which
the decisionmaking apparatus is exogenously given and, for the analyst’s purposes,
fixed, and “theories of institutions” in which some aspects of a larger institutional
structure are endogeneous, in that those who are subject to the rules and impositions

of the institution can in turn influence them. Mayhew (1974, 81-2) puts it well,

10

writing:
[T]he organization of Congress meets remarkably well the electoral needs
of it members. To put it another way, if a group of planners sat down
and tried to design a pair of American national assemblies with the goal
of serving members’ electoral needs year in and year out, they would be

hard pressed to improve on what exists.

Mayhew invites us to view the organization of the US. Congress through the lens of
this analogy, since for him and for generations of later legislative scholars, members
of Congress are very much like this imagined group of institutional planners.

Crucially, though, not all aspects of the decisionmaking apparatus can be so easily
changed as others. While the membership of the US. House of Representatives is free
to remake its internal rules and procedures however it likes, it would be practically
impossible for the House alone to affect a change to the bicameral decision rule
embedded in the constitutional scheme of the American legislative branch. Indeed,
the set of institutional structures that can reasonably be thought of as endogenous
would be a crucial ‘moving part’ in any theory of legislatures that purports to be
broadly generalizable.

Finally, a theory of legislatures must posit the means by which members’ pref-
erences interact with the institutional structure to achieve their goals. At the same
time, we need an account of where those preferences come from. Typically, the pref-
erences of legislators are thought to be induced by constituents, or more broadly
by structural aspects of “the electoral conditions which satisfy [politician’s goals]”
(Schlesinger 1966, 5). While these two conceptualizations might seem to be the same
thing, a general theory must account for the fact that the achievement of politicians’
goals is not always contingent upon the demands of constituents in any direct way.
Carey and Shugart (1995) present a rank ordering of national electoral systems ac-

cording to the relative importance for a legislator’s electoral fate of votes for him

11

(personal votes) and votes for his party (party votes).8

In the United States, where elections are characterized as profoundly candidate-
centered (Mayhew 1974), the electoral fate of a member of Congress is definitively in
the hands of her constituents, and that is where her focus almost always is.9 In an
otherwise similar electoral system like that found in the United Kingdom, however,
electoral success is largely determined by one’s party. Since the parties have signif-
icant control over ballot access and determine which candidates will run in which
district independent of any residency requirement, the most important factors de-
termining the electoral success of the individual members of parliament are 1) the
popularity of their party"), and 2) their favorability in the eyes of the party deci-
sionmakers who determine who runs where (Carey and Shugart 1995). In this case,
despite a single-member district plurality vote system that is otherwise structured
very similarly to the United States system, we end up with legislature in which mem-
bers’ preferences are largely induced by their party and not their constituents, which
is, again, very different from the contemporary understanding of induced preferences
in the study of the US. Congress.

So while thinking in terms of induced preferences should still be an important
assumption for a unified theory of legislatures, we must recognize that variation in
the electoral conditions of a political system profoundly influences who is inducing
those preferences.11 But as to the question of how these preferences interact with the

institutional structures to help members achieve their goals, we have a few ready—

 

8Taylor (2006) makes the connection between the personal vote nature of an electoral system
and the procedural environment in the legislature/parliament, finding intriguing, but still limited,
evidence of a relationship.

9“What is important to each congressman, and vitally so, is that he be free to take positions
that will serve his advantage” [emphasis added] (Mayhew 1974, 101).

1"It is in this way that constituents can indirectly influence their representatives in the British
system.

11At the same time, it is necessary to consider the possibility that legislators’ preferences are
in part ‘induced’ by their own personal beliefs about what makes for good public policy. This
is the position staked out by Fenno (1973) and Rohde (1994), among others. But even a theory
which purports to unify theories of legislatures may have to set aside this possibility for the sake
of parsimony.

12

made answers provided by Mayhew (1974).

For Mayhew, members of a legislature have their preferences induced by the
preferences of their constituents, and use the institutions of the legislature to en-
gage in activities that allow them to 1) make sure their constituents know who they
are (advertising), 2) to know where they stand on issues / question important to the
constituents (position taking), and 3) to make sure constituents know what their
representative has done for them (credit claiming). These activities can be general-
ized beyond the representative-constituent relationship, but the reason that all three
activities are framed in these terms is that Mayhew is building upon his theorized

electoral connection—that members are single-minded seekers of reelection.

1.3 Schlesinger’s Ambition Theory

Schlesinger viewed his theory of ambition as an extension of theories of parties. The
innovation that ambition theory represents is in viewing party organizations as ex-
isting primarily to aid the individual politician in winning and retaining political
office. Schlesinger identified three distinct manifestations of political ambition: 1)
static ambition, in which politicians seek to win and hold a particular office such that
they make a career out of holding it, 2) progressive ambition, in which a politician
views their current office as a means to winning some higher office, and 3) discrete
ambition, in which a politician views their office as a short-term commitment, mo-
tivated by a sense of public duty or by a particular and limited set of public policy
goals.12 Others, notably Aldrich (1995), have built on this foundation, arguing that
party structures within government and without arise to meet the needs of ambi-

tious politicians. Schlesinger is more agnostic about the direction of causality here,

 

12For my purposes, discrete ambition is relatively unimportant except in cases where it is forced
upon legislators by circumstance-notably in the case of term limits. Whether discretely ambitious
politicians are motivated by a sense of civic duty or by policy concerns, it is difficult to make
predictions about that behavior since they lack the representative relationship to a constituency
that allows us to assess their induced preferences. I therefore set them aside, for the most part.

13

though, in that his analysis of party structures and career paths across the American
states reveals wide variation, raising the question of whether the types of ambitions
and the sorts of career paths that we see produce party structures, or whether party
and other structural factors determine the career paths.

Schlesinger was primarily interested in mapping out career paths in the Amer-
ican political system, and in that way he can be seen as engaging in an inductive
process. For example, a fundamental aspect of his analysis dealt with identifying
what he called manifest offices. This refers to the office in which a politician usually
serves when she targets another loffice (in which ambition manifests itself). So when
looking for likely candidates for the House of Representatives, we usually look in
state legislatures, and when looking for likely US. Senate candidates, we often look
in the House, etc. When the literature on congressional elections refers to quality
challengers13 then, this idea of manifest offices becomes important. While challenger
quality usually refers to the holding of any prior political office, there are certain pat-
terns to the typical career path that have important consequences both for elections
and for subsequent policy making. As Rohde (1979) points out, the sorts of quality
challengers we see are driven in part by the fact that the districts /constituencies of
manifest offices are almost always nested within the districts of the higher offices to
which politicians aspire. This can influence the style of campaigning we see (Fenno
1982), but it can also affect how decisions are made in both offices, by incumbents
and by their potential challengers.

It is in this way that Schlesinger’s ideas about ambition can be put to a different
use than he intended. Rather than using ambition theory to explain just the char-
acteristics of parties as organizations, we can use also use it to explain behavior in

and even the structure of legislatures.

 

13For example, see Bianco (1984); Bond, Covington, and Fleisher (1985); Collie (1981); Fowler
(1993); Jacobson (1989); Krasno (1994); Krasno and Green (1988); Lublin (1984).

14

 

1.4 Ambition Theory and the Study of Legisla-
tures

There are two fundamental lines of research in the study of legislatures, especially
as regards the study of the United States Congress. One line of research deals with
the behavior of strategic candidates (Black 1977; Rohde 1979; Jacobson and Kernel]
1983; Jacobson 1987, 2004; Stone and Maisel 2003; Stone, Maisel, and Maestas 2004).
In this literature the focus is on the strategic entry and exit decisions of candidates as
well as the determinants and consequences of candidate quality. Progressive ambition
is a central theoretical concept in this line of work.

The other line of research in the congressional literature deals with the inter-
nal organization of Congress as an institution (Fenno 1973; Mayhew 1974; Cooper,
Brady, and Hurley 1977; Shepsle and Weingast 1987; Weingast and Marshall 1988;
Rohde 1991; Cox and McCubbins 1993; Krehbiel 1991, 1998). The fundamental
focus of the literature on congressional organization, at the very least since May-
hew (1974), has been on how the institutional structures of Congress respond to
members’ desire for reelection. Theories since proposed have attempted to add nu-
ance to this understanding of Congress as an institution. Some have argued for the
importance of accounting for party structures in the organization of congress (Ro-
hde 1991; Cox and McCubbins 1993, 2002, 2005; Aldrich and Rohde 2002). Others
have contended instead that policy preference of the members subsumes party (Kre—
hbiel 1993, 1998; Schickler and Rich 1997) and that the institutional structures of
Congress are designed to provide the information that is used to translate preferences
into substantive policy (Gilligan and Krehbiel 1990; Krehbiel 1991).

What is not in dispute among these scholars is the fundamental importance of
reelection as the motivator for congressional behavior and structure. The closest
we come in the congressional literature to modifying that assumption is in allowing

that members’ individual ideas about what makes good public policy may be an

15

additional factor in explaining what they do (Fenno 1973; Rohde 1994).

The principal contribution of Mayhew’s (1974) groundbreaking theory is the ar-
gument that congressional organization is endogenous; that, in other words, the
institution was created and continues to be shaped by the individuals who oper-
ate within it. Starting with the simplifying assumption that members of Congress
are motivated solely by the desire for reelection, he contends that all aspects of
Congress as an institution can be viewed as serving the reelection motive. This can
be addressed by simply acknowledging that in order for other goals to be achieved,
reelection is the first and most important. In contrast, Fenno’s (1973) analysis of
congressional committees identifies four14 motivations which drive members’ choice
of committees: reelection, power and prestige within the chamber, good public pol-
icy, and the desire for higher 0fl‘lce.’5 It must be remembered, however, as Fenno
himself admits, that the ambition to attain higher office is treated peripherally in his
analysis of committees (1973, 1).16

Mayhew’s argument, put simply, is that Congress is organized in such a way as
to best facilitate members’ pursuit of their interests. If we follow Mayhew’s logic,
for the moment, beyond the reelection motivation, we might ask to what extent
the organizational structure of Congress facilitates members’ pursuit of the goal of
attaining higher office. The literature on congressional organization, however, has
remained centered on static ambition as the fundamental influence on the behavior
of members within the legislature (Schlesinger 1966).

Beginning with Schlesinger (1966), scholars have focused with increasing theoret-

ical and empirical rigor on the behavior associated with the concept of progressive

ambition (Black 1972; Rohde 1979; Brace 1984; Gates and Humes 1997). More

 

1"Fenno mentions in passing a fifth goal, personal gain, that plays no role in his analysis.

15Fenno’s list has been further refined by the argument that “intrainstitutional ambition (the
desire for leadership positions) is a unique form of ambition that results in behaviors distinct from
those generated by either progressive or static ambition (Herrick and Moore 1993).

16In fact, Mayhew’s ( 1974) enumeration (and that of countless others) of Fenno’s proposed mo-
tivations drops the desire for higher office altogether (16).

16

rare have been analyses that examine the consequences of progressive ambition for
behavior within the legislature (Hibbing 1986; Jenkins, Crespin, and Carson 2005).
Beyond members’ roll call behavior, however, there has been little attention paid to
how the creation and utilization of congressional institutions is influenced by mem-

bers’ progressive ambition." Herein, I intend to fill this gap in the literature.

1.5 General Assumptions
In this section, I lay out a set of general assumptions that motivate the project.

OGeneral Assumption 1: In a costless environment, all members are pro-

gressively ambitious.

As Rohde (1979) argues, the offer of a costless exchange of a lower office for a
higher would not be turned down by a rational politician. Higher offices entail greater
benefits for members. Yet running for higher office entails costs and risks. If the costs
are too high, then the desire for higher office is offset. It is when the cost and risks
are acceptable that we should expect progressive ambition to emerge (Black 1972;
Rohde 1979). As Kazee (1994) states, “The decision to run results primarily from
a matching of individual ambition and the context of the opportunities available to
the potential candidate”.

General Assumption 1 does not suggest that all members of the House of Rep—
resentatives value a Senate seat equally, or a Senator the White House, nor does it
suggest that the willingness and ability to bear the costs and risks associated with
running for higher office are uniformly distributed.

To wit,

OGeneral Assumption 2: Difierences in revealed progressive behavior are

 

' 1'ISee Griffin (2003) for the only work, to the best of my knowledge, which touches on this idea.

17

conditional on the set of costs faced by a member and on the willingness

of the member to risk bearing them.

The utility of gaining higher office most likely varies between members, as can the
willingness of members to bear risk (Rohde 1979). Of course, risk bearing as a concept
is fundamental to the study of progressive ambition, but it is doubly important
in building a theory which tries to bring ambition theory to our understanding of
legislative organization. This is because creating such a theory forces the scholar
to make a clear choice in building the argument. The foundation of the argument
could take one of two forms. In one conceptualization, the member chooses to utilize
an institutional structure in a particular manner because of a preexisting, relatively
greater desire to seek higher office. On the other hand, the argument could hold that
certain institutional structures better facilitate the inherent desire of all members to
pursue an office higher up the career ladder while being less conducive to the pursuit
of reelection, meaning that the utilization of such structures is a risky strategy.

Both formulations seem plausible, but I choose to build primarily upon the latter.
The variation among members in their willingness to bear risk is the critical link that
allows the partial reconciliation of the two points of view. If we play out the logic of
the argument, the possibility for that reconciliation becomes more clear. The view
advanced here essentially holds that the utility of attaining a higher office is constant
across members, leaving variables relating to risk-bearing, institutional utilization,
and electoral context to explain the behavioral expressions of progressive ambition.

There is likely to be observational equivalence between this model and that which
allows for the utility of a higher office to vary across members. In fact, the alternative
conceptualization would probably entail subsuming the willingness to bear risk under
a broader conception of the subjective utility of attaining higher office. Thinking
ahead to the stage in the research process when we move from concept to measure,

however, the problem with this alternative is clear. It is quite likely that there is

18

variation in the subjective utility members assign to a higher office. But how would
we go about measuring that variation?

Holding the utility of higher office constant but allowing risk-bearing and the
institutional and electoral context to vary gives us a theoretical model to which
empirical investigation is presumably more applicable, as the crucial link of risk-
bearing offers itself up to measurement much more easily than would variation in
the utility of higher office. This helps avoid a serious problem. In the case where
a model is proposed to explain a member’s decision to run for higher office, we are
predicting the behavioral expression of the underlying latent variable of progressive
ambition. If we were to try and use some measure of the subjective utility of higher
office (which is synonymous with a member’s progressive ambition), the explanatory
side of the equation has been polluted with elements of that which we are trying to
predict.

Instead, we can follow the example of economists of both the neoclassical and
neoinstitutional schools and assume that tastes are exogenously given. In order to
delve into the taste for public office in general, or for specific offices in particular,
we would need to focus on the psychology of politicians. The best opportunity for
doing so would probably employ the theory of achievement motivation, pioneered
by McClelland et al. (1953), and elaborated in important (and politically relevant)
ways by Raynor (1969, 1970) and by Winter (1973a, 1973b). The literature that
followed was principally concerned with academic and occupational achievement,
but the basic theoretical premises could easily be applied to the political setting.
As Arkes and Carske (1982) summarize, the purpose of the theory of achievement
motivation is to explain “individual differences in the motive to enter an achievement
situation” (252). As a theoretical enterprise it is profoundly useful to examine the
psychology of political motivation, but the practicalities of political research make it

difficult to test these theories, since traditional experimental or even survey research

19

is often difficult to employ with politicians as subjects.18

The context in which a member decides also conditions the costs faced when con-
sidering a run for higher office. For example, the presence or absence of an incumbent
(Erikson 1971; Cover 1977; Cover and Mayhew 1977) is a crucial conditioning factor
in the decision to run. But these are factors external to the legislature itself. My
theory holds that conditions internal to the legislature can influence the costs, and

perhaps the benefits, associated with running for higher office.

Thus,

oGeneral Assumption 3: Certain structures within a legislature better

facilitate the pursuit of some goals, but hinder the pursuit of others.

Returning to Fenno’s enumeration of member’s goals, there are clear and meaning-
ful tradeoffs in the choices members make when prioritizing the goals they pursue.
Prioritizing the pursuit of higher office over all else might make it more difficult to
pursue power and prestige within the chamber, or even to win reelection. Presum—
ably, the priorities assigned by a member to each of Fenno’s goals are revealed by
how that member makes use of available institutional resources.

Party as an institutional structure, for example, is likely utilized by members in
various and distinct ways to pursue all of Fenno’s goals. Members can use party
leadership positions to gain power and prestige in the chamber, or trade support
for the party position on legislation for help in getting reelected, or use the party’s
agenda tools to pursue preferred public policy. And returning to Fenno’s neglected
fourth member goal, members presumably use party to pursue higher office. Such
that party, or more specifically party majority status, makes legislative activity less
costly, we would expect that members would use the resources afforded them by the

party to pursue legislative strategies that will aid in reaching higher office (Victor

 

18Though Winter (1980, 1982) and Winter and Stewart (1977, 1978), among others, have at-
tempted to adapt motivational research to the context of political elites.

20

2005). We might also expect that members use their roll call behavior to ‘get right’
with the party in anticipation of running for higher office (Jenkins, Crespin, and
Carson 2005).19 Finally, as will be argued further in the third chapter, members
benefit from the party reputation that is in part derived from the party’s performance
in the legislature, even when they run for a different office.20 Progressively ambitious
members thus have incentives to ensure the protection of the collective good of party
reputation even though they intend to move on to a different office.

Party is not, however, the only institutional structure members use to pursue
their goals. The congressional committee is another structure that can help to fa-
cilitate the pursuit of particular goals, both through explicit design and through
member utilization. The substantive chapters of this dissertation will be bookended
by arguments and analyses about legislative committees, with Chapter 2 illustrating
what ambition theory can teach us about the utilization of committees and Chapter
4 illustrating how the design of committees can be better understood by bringing

ambition theory to bear on the problem.

1.6 Testing Theories of Legislatures

The difficulty in testing theories of legislative organization is that we are not often
present at their creation and, moreover, we cannot impose the sorts of experimental
manipulations on the creation and design of legislatures that would enable us to

put such theories to the strongest possible test. Instead, we must engage in what

 

19Alternatively, the limitations in terms of time and other resources faced by a House member
gearing up for a Senate run may prohibit the construction of an independent legislative program.
Party signals then come to represent a low cost information shortcut for these members, who
still must uphold their duties as members of the House despite the demands of their Senatorial,
gubernatorial, or presidential campaigns. Thus, we would see an increase in party voting loyalty
among these members without any explicit or implicit trading of votes for party support. An even
simpler argument, and one that I expand upon below, holds that members support (or are cajoled
into supporting) the party because its reputation represents a common good which benefits them
regardless of what office they are seeking.

20e.g. The upper chamber in a bicameral system, state/regional executive positions. national
executive posts.

21

King, Keohane, and Verba (1994) call descriptive inference. They write, “Descriptive
inference is the process of understanding an unobserved phenomenon on the basis of
a set of observation” (55).

Institutional change, while not necessarily unobservable, is often difficult to ob-
serve. Even changes to what scholars consider the most fundamental institutions of
the United States Congress have left the basic structure largely intact. Despite its
apparent continuity, many scholars would consider the US. House after Reed’s Rules
to be a very different institution than what had existed prior (Hasbrouck 1927; Cox
and McCubbins 1993, 2005). Similarly, Riker (1955), Rohde (1991), and Katz and
Sala (1996) all point to critical changes to the institutions of Congress that left its
tangible internal structures relatively unchanged. Indeed, Davidson (1990) and Adler
(2002) suggest that changes explicitly aimed at the tangible structures of Congress
(especially reforms to the committee system) are not nearly so effectual as we might
expect.

So how are theories of legislative organization to be tested, if the clearest examples
of legislative change are not considered the most important changes? Our theories
must identify the equivalent of descriptive inferences, often inferences about the
behavioral response of legislators to the institutional factors that drive our theories.
Even when tangible structures remain relatively stable over time, we can assess
changes in institutions by looking for changes in behavior. Indeed, fundamental
theories of economic exchange (and of exchange under different institutions), which
so crucially inform scholars’ understanding of legislatures, are often tested using the
behavioral response of market actors, i.e. the price of goods that are bought and
sold.

First, we can test behavioral response cross-sectionally, examining whether indi-
viduals are behaving the way our understanding of the institution predicts. In this

manner, we make predictions about behavior that are derived from an understanding

22

of the institution. To preview the arguments I will advance along these lines, I spend
next three chapters making arguments about the significance of career goals beyond
reelection in shaping organization, and testing them in a number of contexts. In
Chapters 2 and 3, I use this cross-sectional approach, focusing on committees and
parties. In Chapter 2, I argue that certain committees facilitate activities that are
necessary for the pursuit of higher office. If organization does not respond to the
desire of some members of a legislature to run for higher office, then we should expect
no difference among committees in the propensity of their members to run for higher
office. Evidence of a difference, which in isolation is just a behavioral response to the
institution, can be seen as evidence in support of a theory of legislative organization
that has a place for a broader understanding of members’ goals.

In Chapter 3, I argue that while running for higher office implies the severing of
a member’s connection with the party in the legislative chamber, the electoral value
of the party’s reputation remains valuable to the departing member, such that party
demands on procedural votes are still binding on the member. Finding evidence in
support of the argument, I would point out that the argument is not just about how
higher oflice ambition influences behavior, but it also tells us something we did not
know previously about the delegation relationship between members of the United
States House of Representatives and their parties.

Party reputation appears valuable enough to members that even when they are
immune to an important subset of the rewards and punishments that parties use to
maintain discipline on procedural matters in the House (relating to committee posi—
tions, floor access for legislation, party leadership posts, etc.), progressively ambitious
members still acquiesce to these party demands. This has important implications
for partisan theories of legislative organization, which rely heavily on the concept of
party reputation to motivate their predictions about structure and behavior (Cox

and McCubbins 1993, 2005).

23

Even better than taking such a cross—sectional view, as I (in essence) do in Chap-
ters 2 and 3, we can examine behavior longitudinally, assessing behavioral response
to critical events that have important institutional consequences, according to the-
ory. To paraphrase Field (1981, 183-4), as the underlying parameters of a political
system change, rules, originally efficient, become inefficient. Rule change in the leg-
islature results when these underlying parameters influence the membership and the
operation of the legislature. In Chapter 4, I look at an important example of change
in these underlying parameters, the imposition of legislative term limits, and derive
predictions about particular institutional changes in response to term limits.

In either approach, we should be looking for how the preferences, interests, and
goals of legislators affect institutional structure. That we look for evidence of these
sorts of institutional arguments in the behavioral consequences thereof is not an ac-
cident, as behavior is often much easier to measure, and many of the most important
institutional features of legislatures are less than tangible, and thus much harder to
measure than might otherwise be. This is not a problem unique to the arguments I
advance in this dissertation. In fact, this is the approach that most of the literature
on legislative organization has been forced to employ for decades.

In the late- 19808, there was a significant debate in the congressional organiza-
tion literature about the power of committees. Built largely as a justification for
Shepsle’s (1979) committee-based structure induced equilibrium solution, Shepsle
and Weingast (1987), Weingast and Marshall (1988), and Gilligan, Marshall, and
Weingast (1989) advocated an understanding of Congress that embedded powerful,
agenda-controlling committees in an understanding of congressional institutions. The
upshot of these arguments is that committees get what they want in terms of leg-
islative outcomes. But the implications of these arguments that are actually tested
relate to behavior, and specifically relate to which members sit on which standing

committees, and on which members end up on the conference committees that help

24

determine the final form of legislation after it has passed both chambers of Congress.

The critical prediction is that the preferences of standing committee and con—
ference committee members should be unrepresentative of the legislative chamber
as a whole. Critics of these arguments fight on the same ground, pointing out that
committees are largely representative of the chamber, and that floor-based majorities
have the means to circumvent committee power even when committees are unrep-
resentative (Krehbiel 1987, 1991). Even though the most direct implication of the
institutional arguments put forward by these scholars might be legislative outcomes,
they content themselves with finding more easily measurable, but at the same time
more indirect, theoretical implications.

Moving into the 1990s, a similar debate arose regarding arguments that posited
a strong, agenda-controlling majority party in the House of Representatives (Cox
and McCubbins 1993). Again, while the fundamental argument was about how
the majority party monopolizes the power that in principle resides in the whole
chamber and uses that power to bias the operation of the institution is ways that
favor its members. One would expect tests of the argument to relate to legislative
outcomes, since partisan and nonpartisan theories of legislative organization have
clear, divergent predictions about outcomes: either they are biased in favor of the
majority party or they are not. But legislative outcomes are difficult to measure,
as they are at their base the result, jointly, of a proposal and a status quo being
targeted by the proposal, neither of which are easily measurable.

Instead, partisan theories (again exemplified by Cox and McCubbins 1993) test
implications of their theory that again relate to committee membership. So is the
membership of committees representative of the preferences of the chamber as a
whole, or of the majority party membership? And are the committees with the
most influence over the agenda and with the most potential for creating externalities

(positive or negative) dominated by majority party interests? Again, Krehbiel (1993)

25

responds, constructing a convincing argument that partisan theories have important
implications for roll call behavior (perhaps a more direct implication of the theory),
but that these implications are not those tested by advocates of partisan theories.
Despite the aptness of his criticism, however, Krehbiel (1993) goes on to subject
partisan hypotheses to tests relating to standing committee and conference committee
membership.

This is not to say that these are inappropriate tests of theory. The difficulty in
social science, as King, Keohane, and Verba (1994) tell us, is that we so often deal
with theoretical concepts that are inherently qualitative (e.g. parties ‘matter’ in the
legislature, or committees are powerful) and must identify observable implications,
often also qualitative, but that lend themselves to quantification. We must often
look, then, for the shadow of the institutional effects about which we construct our
theories. Our interest may be in asserting that X is true, but if X is not directly
observable, then we identify an observable phenomenon Y that follows from X.
Finding evidence of Y then lends support to the theoretical assertion of X. The
strength of the inference we make in such a case is based both on how strong the
case is that Y follows from X, and on how convinced we are that Y is not being
caused by some other unobserved factor. Good theory and clever research design
make such inferences much easier.

There are examples in the literature on legislative organization, of course, where
scholars test more direct implications of theory, and do so longitudinally with an with
an eye toward identifying critical events that have consequences for theory. Examples
include analyses of rules changes in Congress over time (Binder 1997; Dion 1997), the
rise of committee seat property rights in the House (Katz and Sale. 1996), and the rise
of a more powerful majority party through changes in internal caucus rules (Rohde
1991), all of which put forward arguments about how legislatures are structured,

and identify particular events and factors that predict observable changes in the

26

institution. It is this sort of analysis, moving from good theory to good research
design, which enables us to more confidently claim empirical support for particular
institutional arguments.

Again, I do not claim that the work I present here avoids the pitfall of testing
indirect implications of theory, nor do I claim that the research designs I employ
are the ideal. Instead, I think it is useful to point out that many of the admitted
shortcomings of this dissertation are common to the study of legislative organization,
and perhaps to social science more generally. Crucially, scholars must recognize the
shortcomings of their analyses and be up front about them, and at the same time
work towards tests of theory that focus on more direct implications of theories, and
build research designs that allow for more confident inferences about the empirical

support found by analyzing the shadows of institutions.

1.7 Competing Theories: What Ambition Con-
tributes

The construction of critical tests mirrors the broader evaluation of competing the-
ories. What should we expect to see if a theory is true? What is the appropriate
null against which we should compare our theory and its predictions? The prob-
lem for an ambition theory of legislative organization is that there is no existing,
well-formulated theory with which to juxtapose it. The theoretical disagreements in
the literature tend to focus on a subset of the theoretical space, so to speak. Are
committees independently powerful of the parties and/ or the center of opinion in
the chamber? Are party and induced preference separable factors in explaining the
structure of the Congress, the behavior of members in it, and the policy outputs it
produces? What is the extent of agenda control in the Congress, and who exercises

it? Do the similarities and differences between the House and Senate matter, in

27

terms of agenda setting and policy choice?

All these questions have the assumption of the reelection imperative in com-
mon, and so jointly represent a general perspective against which ambition theory
of legislative organization can compete. This is (with some exceptions) the minimal
goals perspective advanced by Mayhew (1974), that members are driven solely by
the desire for reelection, which is, after all, a necessary condition for the achieve-
ment of many other goals we might posit for elective politicians. This was precisely
the argument Mayhew used in setting up his theoretical disagreement with Fenno
(1973), who posited a more expansive set of goals for members of Congress. Fenno’s
behavioral assumptions subsume the minimal goals perspective and the ambition
theory perspective, since Fenno posited reelection, progressive ambition, concern for
good public policy, desire for power and prestige within the chamber, and pure self-
ish interest as potentially important goals for members of Congress. Another set
of questions arises from Fenno’s (1973) perspective. What is the significance of the
different delegations of power and authority members of Congress make? Are the
priorities of party and committee shaped by members’ personal policy preferences,
rather than their induced preferences? Why can the Congress sometimes make policy
that promotes the general welfare, and why not other times?

The key to understanding the relationship between the minimal goals approach
of Mayhew, the ambition theory of legislative organization, and the expansive goals
perspective of Fenno is to recognize that they are not tailored to answering any of
the particular questions I have just listed, but that they are general perspectives that
imply particular questions. These theories represent a hierarchy of complexity, with
each theory making competing claims about the structure and behavior we should
expect to observe in legislatures.

But how should we assess these theories? Some questions are not implied by each

of the theories. Some of the questions cannot be addressed at all by the minimal

28

goals perspective, except in its role as the null against which alternative predictions
will be tested. The key to assessing the relative merits of these theories is to find the
phenomena of interest they share in common, to derive their respective predictions
about those phenomena, and to test them empirically. For example, we can derive
divergent predictions about the structure of the committee system in a legislature
from each of these perspectives.

The minimal goals perspective suggests two notions about how the committee
system should be constructed. One, which is less realistic in the context of the
US. Congress, suggests that there is an optimal mix of activities and policy areas
for all members, and so committees should be homogenous in their construction.21
Alternatively, this perspective suggests that the variation in subject matter and in
the mix of activities among committees should correspond to the variation in the
nature of members’ constituencies.

The ambition theory perspective builds on that second notion, adding that there
should be further heterogeneity in committees vis-a-vis the mix of activities they
offer that is independent of constituency characteristics, corresponding to the differ-
ent sorts of ambition held by members. Fenno’s more expansive perspective predicts
heterogeneity among committees in the ways already discussed, but adds that com-
mittees should vary in their relative prestige, that committees should offer positions
to members on the basis of their personal interests in particular policy areas, and
even (perhaps) that some committees should offer opportunities for graft while in
office or for transition into lucrative positions in the private sector after leaving office.

As usual, the burden of more complex theories is to explain enough more than
simpler theories to justify the added complexity. The ambition theory perspective

is between the other two perspectives in terms of its complexity, but is closer in

 

21The so-called standing committees of the British House of Commons could be interpreted this
way, as they have no subject matter jurisdiction and no permanent membership, merely taking
turns considering bills with members volunteering, taking turns, or being pressed to serve during
the consideration of a bill.

29

spirit to Mayhew’s minimalist perspective. This dissertation puts to the test only
a very small portion of the important predictions that could be derived from this
perspective. But as a perspective, I think it has important implications for the study

of Congress and of the construction of more general theories of legislatures.

1 .8 Organization

The rest of the dissertation is organized as follows: In Chapter 2, I examine how leg-
islators use committees to aid them in their pursuit of higher office. More specifically,
I look at the United States House of Representatives and examine the relationship
between membership on committees that facilitate certain activities is related to re-
vealed ambition for a Senate seat. In Chapter 3, I present an argument dealing with
the role parties play in members’ strategies for seeking higher office. I argue that
what is typically understood to motivate reelection-minded members of the House
to support their party on procedural matters also applies members seeking higher
office; namely, party reputation. In Chapter 4, I address how an ambition theory of
legislative organization yields predictions about the endogeneity of institutions that
are impossible using contemporary legislative theories drawn from the study of the
US. Congress. More specifically, I present an argument about the maintenance of
both individual and collective committee property rights, and how our understand-
ing of these priorities is shaped by the assumptions we make about legislators’ career
goals. I argue that state legislative term limits tend to create pressure to ‘devalue’
committee seats and committee jurisdictions. Chapter 5 presents a summary of the
dissertation, along with some concluding remarks about what I have learned in the

process of writing the dissertation.

30

Chapter 2

Progressive Ambition and
Committee Membership in the
House of Representatives

2. 1 Introduction

In this chapter, I propose a theory that views progressive ambition as an important
factor shaping the utilization and potentially the design of the internal structures of
Congress. I focus on the House Judiciary Committee as an example of a committee-
as-‘incubator’ for progressively ambitious members. I present findings which provide
supporting evidence for the contention that members of the Judiciary Committee
should be more likely to run for Senate than their counterparts on non-exclusive,
non-control committees.

The story that informs this chapter about the House Judiciary Committee and
its relationship to higher office aspirations could be told in several ways, but I would
like to begin, strangely enough, in a different committee entirely. A wonderful il-
lustration, though, of the theoretical proposition that underlies my argument—that
certain committees better facilitate the activities associated with running for higher
office—can be found in a committee called, at various times during its existence, the

House Committee on Un—American Activities1 or the House Committee on Internal

 

lOften erroneously referred to as HUAC, presumably to facilitate pronunciation.

31

Security.

The story (which will eventually bring us around to the House Judiciary Com-
mittee) starts there not just because the jurisdiction of Internal Security was rolled
into the Judiciary Committee in 1975, but also because HUAC/ Internal Security,
even more than Judiciary, was archetypical of the sort of committee that facilitates
members’ pursuit of higher office. As I will discuss in more detail below, advertising
and position taking are the most useful of Mayhew’s (1974) enumerated activities for
the higher office aspirant. The story with which I introduce the theory and analysis
of this essay, then, is an effort to show how members of Congress can use particular

committees to engage in these activities.

2.2 A Case Study in the Use of Committee Mem-
bership to Pursue Higher Office

Committee membership can be electorally useful in a number of different
ways. Some committees supply good platforms for position taking. The
best example over the years is probably the House Un—American Activi-
ties Committee ([later] the Internal Security Committee), whose members
have displayed hardly a trace of interest in legislation... (Mayhew 1974,
85).

To narrow the scope of the anecdotal evidence provided by this story, I focus on a
rather notable alumnus of HUAC / Internal Security—Richard Nixon. While it would
be a profound error to draw inferences from Nixon’s career as a member of the
House of Representatives and to try to generalize to about all House members, an
understanding of his time in the House can give us insights into the causal mechanism
that underlies the argument I put forward below. Specifically, Nixon’s story typifies
how members of congress engage in position taking and advertising, and how they

use the committee system to do so, since he was one of the best at doing just that.

32

It was the Hiss case which involved espionage charges against a highly—

placed State Department official which put [HUAC] on the front pages

and won notoriety and political prominence for its newest member, Richard
M. Nixon (R—CA). Nixon parlayed the publicity into a US. Senate seat

and the Vice Presidency within a four-year span [emphasis added] (Nel-

son 1994a, 1006).

In his memoir, Nixon claimed that his membership on HUAC was an afterthought;
that, in fact, he only joined the committee as a favor to incoming Speaker of the House
Joseph Martin (R—MA). Nixon’s first choice of committee was Judiciary, on which
he did not win a seat. His second choice was the Education and Labor Committee,
on which he did receive a seat, a choice largely driven by his antipathy toward the
Congress of Industrial Organizations (CIO) due, in part, to its activities on behalf
of his incumbent opponent in the 1946 campaign, H.J. “Jerry” Voorhis (D-CA).

At the same time, membership on Education and Labor put Nixon in a position
to help push labor legislation that was a high priority on the Republican majority’s

agenda for the 80th Congress.

The Republicans’ first priority was to amend, if they could not kill out-
right, the National Labor Relations Act of 1935, which they regarded as
practically the birth certificate of the CIO...Nixon wanted to be in on the
attack, at the heart of the action, and immediately after his election he
had set about getting a place for himself on the Education and Labor
Committee (Ambrose 1987, 142).

As for his placement on HUAC, though, Nixon writes,

Most freshmen received only one committee assignment, but Joe Mar-
tin, the new Republican Speaker, asked if I would be willing to sit on the

House Committee on Un-American Activities. Now that the Republicans

33

controlled Congress,we would be held accountable for the committee’s
frequently irresponsible conduct. “We need a young lawyer on that com-
mittee to smarten it up,” he said, and he added that he would consider
my acceptance a personal favor. Put in this way, it was an offer I could
not very well refuse. I accepted with considerable reluctance, however,
because of the dubious reputation the committee had acquired under
its former chairman, Martin Dies, a flamboyant and at times demagogic

Texas Democrat (Nixon 1990, 44-5).

This despite the fact, as biographer Stephen Ambrose points out, that HUAC was a
very high demand committee for Republicans in the 80th Congress. Ambrose argues
further that sitting on Education and Labor and on HUAC “put Nixon at the point
position on the two most important issues on the Republican agenda, curbing the
unions and rooting out the subversives” [emphasis added] (1987, 143).2

Nixon suggested that membership on HUAC was politically risky, but called it
“an opportunity as well as a risk” (Ambrose 1987, 143). Nixon’s administrative
assistant, William Arnold, later wrote that his choice of committees as a freshman
was indicative of Nixon’s “characteristic prescience” (Arnold 1975, 7). Despite his
stated priority with respect to his committee seats, though, Nixon’s first floor speech
in the House of Representatives, on February 18, 1947, involved his work on HUAC
and not his work on Education and Labor, when he made the case fora resolution
citing alleged communist subversive Gerhart Eisler with contempt of Congress for
refusing to testify before HUAC (Goodman 1968, 190-1).

A little over a year later, in May 1948, HR. 5852, known also as the Subversive
Activities Control Bill passed the House, 322-61.3 The bill, also known as the Mundt-

 

2Nixon’s gaining two committee seats is also suggestive of a strategy I discuss later in this chapter
and in Chapter 5, which I call ‘portfolio building,’ in which legislators endeavor to demonstrate
familiarity with and competence in many policy areas by serving on multiple committees and, more
generally, by participating in a broad set of policy areas.

3See Keith Poole’s Voteview program, available at http: //www.voteview. com/ for further in-
formation on this roll call vote, # 127 of the 80th House.

34

Nixon Bill, failed to make any headway in the Senate, but represented a rather
high-profile success for the relatively junior Republicans from California and South
Dakota, respectively. If passed, the bill would have required the registration of
individual communists and communist organizations and would have barred known
subversives from holding public office.4 Most of the bill’s provisions would go on to
pass the Senate in 1950 and finally become law as the Internal Security Act, also
known as the McCarran Act (Goodman 1968, 291). In its initial consideration in
1948, Nixon writes, “[Mundt—Nixon] was the first piece of legislation to emerge in
ten years from the House Committee on Un—American Activities” (Nixon 1990, 46).
In the meantime, however, an even bigger opportunity for advertising and position
taking came along for Nixon and the members of HUAC.5

In the summer of 1948, a former communist named Elizabeth Bentley testified
on separate occasions to a federal grand jury in New York, to the Senate Investigat-
ing Committee, and finally to HUAC, about the subversive activities of people she
knew and had heard about from others. Looking for corroboration of her testimony,
HUAC’s chief investigator identified Whittaker Chambers, a journalist working at
Time Magazine, as a potential witness due to his purported ties to the Communist
Party in the late 19203 and early 1930s. Chambers named eight former governing
officials, who had been placed in the various New Deal agencies, the Justice Depart-
ment, and the State Department, as members of the Communist Party and as Soviet
agents (Goodman 1968, 244-252).

For longtime members of the committee, the testimony of Bentley and Chambers

offered vindication for arguments they had been making for years about communist

 

4Cosponsor Karl Mundt, for his part, parlayed the notoriety he received from his HUAC activities
into election as US. Senator from South Dakota in 1948, where Nixon would join him two years
later (Nelson 1994b, 646).

5Goodman (1968, 255) writes of Nixon on the eve of the Hiss case, “Nixon had already shown
himself to be by far the most competent member of the committee. He studied the subject at hand,
paid attention to the testimony, and alone of the nine Congressman was able to lead a witness,
without badgering, along an orderly line of questioning.”

35

infiltration of government and, especially, of the New Deal. But for the commit-
tee’s newest member, Nixon, the hearings surrounding their accusations would allow
him to reinforce the position he staked out with Mundt—Nixon, and even more impor—
tantly, make him a nationally known figure. This opportunity came when one former
official was identified by Chambers: a former government lawyer named Alger Hiss,
who served in several federal departments and agencies, most prominently in the
State Department during the crucial years when the Allied Powers were negotiating
the map of post-World War II Europe and working to establish the United Nations.
In 1948, when Chambers’s accusations were aired the hearing by HUAC, Hiss was
serving as the president of the Carnegie Endowment for International Peace.

Hiss, unlike others who were named by Bentley and Chambers, insisted on appear-
ing before HUAC to rebut the accusation that he had provided secret documents to
Soviet intelligence, crucially choosing not to assert his Fifth Amendment right against
self incrimination. He denied having been a Communist, denied passing intelligence
to the Soviets, and even denied knowing who Chambers was.6. This eventually led
to a day—long hearing later in the summer in which Hiss, Chambers, and Nixon and
the other members of HUAC would play out their drama before the nation.7

In the aftermath of the confrontation, Hiss unsuccessfully sued Chambers for li-

 

6At this point, the Republican members of HUAC despaired, thinking they had aimed too
high and would be humiliated for targeting Hiss. Nixon and HUAC investigator Robert Stripling
convinced the committee to pursue the matter despite Hiss’s vehement denials (Goodman 1968,
255—6)

7I use the term drama here for very specific reasons. It is no accident that one of the most
voluminous studies of the history of HUAC’s hearings, entitled Thirty Years of Treason, was written
by Eric Bentley, a noted drama critic and playwright. One of the insights about position taking
Mayhew (1974) provides, which is later reinforced by Fiorina (1989), is that it is not necessarily
about accomplishing anything. The much more important thing, and the lesson that Nixon’s use
of his seat on HUAC teaches us, is that the primary goal of position taking is to be seen taking the
position. Mayhew (1974) writes

The congressman as position taker is a speaker rather than a doer. The electoral
requirement is not that he make pleasing things happen but that he make pleasing
judgmental statements. The position itself is the political commodity [emphasis added]
(62).

Drama, spectacle, etc. can be and have been useful tools in communicating politicians’ positions
to their targeted audiences.

36

bel, Hiss was twice prosecuted for perjury and was convicted in the second trial, and
Richard Nixon became a national figure as a freshman member of the House. Former
HUAC chairman Martin Dies later wrote, with some exaggeration and with seem-
ing bitterness, that “Nixon was the only Congressman to profit by anti-communist
activity.” (Dies 1963, 179).

Nixon himself judged his involvement in the Hiss case in the context of the debate
among his staff and supporters about whether he should run for the California Senate

seat then held by Democrat Sheridan Downey.8 He writes,

Virtually all my friends and advisers told me that running for the Senate
would be tantamount to political suicide. But I recognized the worth of
the nationwide publicity that the Hiss case had given me—publicity on
the scale that most congressman only dream of achieving. Running for
the Downey seat was the only possibility for me to move up the political

ladder at a time when my political stock was high (Nixon 1990, 72).

What should be taken from this anecdote? First, Nixon served on more than one
committee, which may suggest that the building of a broader legislative portfolio is
an important factor in running for higher office. Second, he seems to have chosen
committees in which he would be well-placed to associate himself with the priorities
of his party, in that making changes to labor law and pursuing subversive elements
in government and in society were fundamental to what congressional Republicans
saw as their policy mandate under their new majority. Third, and most importantly,
Nixon was placed on committees that offered important opportunities for position
taking and advertising, and fewer opportunities for credit-claiming. Perhaps it is
this that led Nixon to worry that a seat on HUAC could be risky. While Nixon

legislated from his seat on HUAC, more than anything he used it as a platform

 

8Downey would later withdraw from the race due to health problems, and was succeeded for the
Democratic nomination for Senate by Helen Gahagan Douglas, whom Nixon would go on to defeat
in the general election.

37

to gain public exposure and to make sure the public knew that he was stridently
anti-communist. And in less than 4 years from having taken his seat in the House,
Nixon won a Senate seat, and two years after that he was the Republican nominee
for vice-president. Nixon may have been unique in his meteoric rise, but the strategy
and tactics he used are available, to some extent, to every member of the House.

I would hesitate, at the same time, to make too much out of the merger of HUAC
into the Judiciary Committee, but it is suggestive of the style of politics in which
members of Judiciary tend to engage. As recently as April 2005 a bill, HR. 1974,
to provide criminal penalties for the desecration of the flag of the United States was
under consideration in the House Judiciary Committee. It was a hopeless bill, both
in terms of its fate within the chamber (it died in subcommittee) and in terms of
its ability to withstand constitutional challenge. But the two initial sponsors, both
members of the Judiciary Committee and, interestingly, both Democrats (Boucher
of Virginia and Schiff of California), got their names associated with an issue over
which there is considerable passion. Whether either runs for higher office in the

future remains to be seen.

2.3 Assumptions & Hypotheses

oAssumption: Some committees in the House can better facilitate the

pursuit of particular goals.

Fenno (1973) argues that members use committee assignments to facilitate the achieve—
ment of their goals. Progressive politicians should therefore use committee assign-
ments to pave, in some way, the road to higher office. To flesh out this supposition, I
begin by conceiving of committees as offering differential mixes of Mayhew’s theorized
three activities: credit claiming, position taking, and advertising. While suggesting

that some committees are designed to provide better opportunities for one or an-

38

other of these particular activities, Mayhew provides no systematic framework. Of
the activities and their relationship with the pursuit of higher office, Mayhew (1974)

writes

Higher aspirations seem to produce...[a] distinctive mix of activities. For
one thing credit claiming is all but useless. It does little good to talk
about the bacon you have brought back to a district you are trying to
abandon...Office advancement seems to require a judicious mix of adver-

tising and position taking (75-6).

When thinking about progressive ambition as a motive for committee choice, then,
a crucial distinction would seem to exist between distributive and non-distributive
committees. The fundamental motivation behind seeking a seat on a distributive
committee would seem to be the facilitation of credit claiming, an activity Mayhew
claims helps little in the pursuit of higher office. And while the ‘control’ committees
are committees that stand out from the rest, this seems principally to be a function
of their use by members to gain power and prestige within the chamber (Fenno 1966,
1973) or of their use by the majority party to control the chamber and its agenda
(Rohde 1991; Cox and McCubbins 1993) rather than as an avenue to become known
to a state- and nationwide audience and to stake out positions on national issues.

Keeping this puzzle in mind, then, we come to the principal theoretical hypothesis

of this chapter:

oHypothesis 1 A: Members of the committees that facilitate position taking
and advertising are more likely to run for higher ofi‘ice than members on

other committees.

Here, I put forward a hypothesis that follows from the contention that a member’s
choice of committee can help lower the cost of engaging in the activities that aid in

the pursuit of a particular goal. But is there an identifiable committee or type of

39

committees that offers the mix of advertising and position taking necessary for the
fulfillment of progressive ambition?

While not stating the matter explicitly in terms of progressive ambition, May-
hew cites the now defunct Un—American Activities Committee as an example of a
committee used (and probably created) for the purpose of position taking (85). But
among the current standing committees, I would suggest the Judiciary committee
as a potential candidate for this categorization. Judiciary has been a committee
that deals with some of the most contentious issues of the day: busing, flag burning,
abortion, impeachment, etc. While this aspect of the committee’s activities could be
viewed as placing it under the category of policy committee (Rohde 1991), I contend
that the contentiousness of the committee’s jurisdiction also makes it a good place
for members to engage in position taking and advertising.9

Shepsle (1978) finds that members’ requests for assignment to the House Judi-
ciary Committee are unrelated to the characteristics of their districts. In fact, the
best predictor of a member request for a seat on Judiciary in Shepsle’s analysis is
whether the member is a lawyer. Past members of the committee have identified it
as being a particularly poor base for fundraising purposes (Perkins 1981). The desire
for membership on this committee could be driven by policy interest (likely) or by
the desire for power and prestige within the chamber (less likely) (Perkins 1981).
Also note that Hypothesis 1A refers to higher office in general terms. For the sake of
simplicity, the focus here is on the pursuit of a Senate seat as the behavioral expres-
sion of progressive ambition, though I think it certainly plausible that the argument
could hold more generally, for other offices up the career ladder. So to phrase the

hypothesis in more specific, empirical terms:

oHypothesis 13: Members of the House Judiciary Committee are more

 

9In fact, the lack of opportunities for credit claiming and “pork-barreling” on the Judiciary
Committee is what makes it a risky choice of committee (and thus an expression of the willingness
to bear risk) for members who are statically ambitious, either in the short or in the long term.

40

likely to run for Senate than members not on the Judiciary Committee.

This choice is somewhat arbitrary, but there are a few reasons why I avoid com-
plicating the hypothesis by introducing runs for governor or president. First, more
members of the House run for Senate than for governor or president. Gubernatorial
elections in many states take place in off-years, meaning that members from those
states can run for governor without forgoing a reelection campaign for their House
seats. This might make such a candidacy less risky than running for Senate, but
at the same time, many governorships are term limited, which could make such a
position less valuable. Given that a governorship is an executive branch position,
demonstration of competence as an executive may require very different sorts of be-
haviors within the House. In fact, there are interesting hypotheses to be generated
by these speculations. But on the whole, the set of factors that would need to be
accounted for to make gubernatorial and senatorial candidacies comparable are be-
yond my reach at this stage. As for presidential candidacies by House members,
there simply are not enough of them to conduct any quantitative analysis, at least

for the sample of observations that I have.

2.4 Analysis

In order to test Hypothesis 1B, I utilize pooled data on committee membership
and Senate candidacies from the lolst through 108th Congresses.10 I include as
observations those representatives in states that held a Senate election during a
given Congress“, yielding an N of 2171.

Over the course of this timeframe, about 3.6 percent of all House members made

 

10Sources for this data come from the Clerk of the US House of Representatives:. http : / / clerk.
house .gov, from the Political Graveyard website: http : //politica1graveyard . com, and from the
Congressional Biographical Directory: http://bioguide.congress.gov/biosearch/biosearch.
asp.

11So for example, since the state of Michigan had no Senate race on the ballot in 2004, and thus
no Michigan House members could have run for Senate, they are excluded from the sample.

41

a run at a Senate seat when one was on the ballot in their state. About 4.5 percent
of Judiciary members ran for Senate over the same period. This is only a slight
difference, but it is in the hypothesized direction, and provides some preliminary
support for the hypothesis. I would hesitate to draw any inference from this difference
without controlling for other relevant variables, and so I now move to a multivariate
test of the hypothesis.

Due to the nature of my research question, logistic regression is used to estimate
our econometric model. I pool the data and treat them as a cross-section, but given
that there are multiple observations of the same members (since the unit of analysis
is the individual member by Congress), I must account for the nonindependence of
observations. I do this by standard errors adjusted for clustering on the member
(for a discussion of clustered standard errors, see Primo, Jacobsmeier, and Milyo
Forthcoming).12

My interest is in progressive ambition as an important factor shaping how mem-
bers interact with the internal structures of Congress. I focus on the House Judiciary
Committee as an example of a committee-as—‘incubator’ for progressively ambitious
members. A dummy variable for membership on the Judiciary Committee is the
principal explanatory variable. Since I expect this committee to facilitate members’
advancement to higher office, the coefficient for this variable is expected to be positive
and significant.

One of the goals of members as stated above is the desire for power and prestige
within the chamber. I first control for power within the chamber by including a
dummy variable for those members who are chairs of the full committee.13 In addition

I control for membership on committees of power. I use a dummy for membership

 

12I have also estimated models using Beck, Katz, and 'Ihcker’s (1998) time—series cross-sectional
technique and using Cox’s Proportional Hazards model with the event defined as a run for Senate,
neither of which produced findings that were substantively different from those produced by the
relatively simpler technique I use here.

13The committees coded were the standing committees from the 101st to the 108th Congresses
and the Intelligence and Homeland Security Select Committees.

42

on the Exclusive and Control Committees (Appropriations, Rules, Ways and Means,
Budget, and Energy and Commerce after 1994).

One of the goals of members as stated above is the desire for power and prestige
within the chamber.14 I control for one manifestation15 of this goal, which should
conflict with the pursuit of higher office, by including a dummy variable for those
members who are committee chairs.16 In addition I control for membership on ‘party
committees’. To measure this, I use a dummy for the Exclusive and Control Commit-
tees (Appropriations, Rules, Ways and Means, Budget, and Energy and Commerce
after 1994).

The intuition for this draws on Cox and McCubbins’ (1993) committee analysis.
These committees are fundamental tools by which the parties (and especially the
majority party) control their caucuses and contend with each other for control of
the chamber. To the extent that membership on these committees represents a
‘vote of confidence’ in the individual member by her party, or even service paid by
the member to the party in exchange for party support in a run for higher office,
I think it important to distinguish these committees from the rest. The empirical
manifestation of my theoretical expectation is the covariation between individual
attributes and behaviors and the subsequent decision to seek higher office. But
service on these party committees potentially links individual members to public
judgments about the collective attributes and ‘behaviors’ of the party (Cox and
McCubbins 1993), which risks muddying the waters for the analysis. Thus to make

 

14It is not clear that the goals of power and prestige within the chamber and the pursuit of
higher office must be directly traded off. The national or statewide exposure afforded a member
who becomes a party leader or committee chair could be a useful resource when contemplating a
run for higher office. Such would, however, also make one’s current office much more valuable, thus
making the expected utility of moving from House to Senate, for example, less great. So again, the
nature of the tradeoff is unclear.

151 have also tried controlling for a rather narrow set of leadership positions (Floor Leader, Party
Whip, and Conference/ Caucus Chair) and found that holding one of these positions perfectly
predicts failure (i.e. not seeking a Senate seat) in the sample.

16The committees coded were the standing committees from the 101st to the 108th Congresses
and the Intelligence and Homeland Security Select Committees.

43

an appropriate comparison, I control for Exclusive and Control Committees to remove
these committees from our reference group.

The decision to seek higher office is the product of a calculation of costs and
benefits (Black 1972; Rohde 1979) that derive from risking a currently held office
for the chance to gain higher office. The costs are represented by the difficulty of
winning the higher office. I also control for these factors. The presence or absence of
an incumbent is important, as is the partisanship and vulnerability of the incumbent.
Dummy variables are constructed for those Senate seats without an incumbent at a
given election, for those with incumbents facing reelection for the first time (fresh-
man), and those with an incumbent of the opposite party.

The electoral circumstances a member may face will directly affect their decision
to run. The past performance of the presidential candidate in a state is a rough clue
as to the level of partisan support that may exist in the state. I measure this as the
percentage of the two—party vote in the previous presidential election received by the
candidate of the member’s party.

With the expectation that the overlap of the district that one holds to the district
one is seeking matters, I record the state population in thousands to account for
the relationship between House districts, which are fixed in size, and the state. I
thereby aim to capture and control for the heterogeneity of a member’s state (Lee
and Oppenheimer 1999).

Finally, I consider the role that policy considerations play in shaping members’
behavior. If a member is motivated by policy concerns, she should seek out commit-
tees that would facilitate those ends to support her ideological goals. Impressionistic
evidence suggests that Judiciary draws members with extreme policy positions. This
policy motivation could plausibly be the factor leading members of Judiciary to sub-
sequently run for higher office, suggesting the potential for a spurious result in an

analysis of the effect of the Judiciary committee variable. Therefore, I use the ab-

44

solute value of members’ first-dimension DW—NOMINATE score as an admittedly
rough measure of ideological extremity.17 I also include party-Congress interactions
(to capture partisan electoral tides) and ICPSR region variables as fixed effects.18
The results of the logit estimation are presented in Table 2.1.19

Hypothesis 1B faces a rather straightforward test. The coefficient for judiciary
committee should be positive and significant, and, as Table 1 shows, it is.20 Interest-
ingly, the same is true of the coefficient for the dummy capturing exclusive and control
committees.21 Members of Judiciary are 2.9 times more likely to run for Senate than
members not on Judiciary.22 Members of the exclusive and control committees are
estimated to be about 2.7 times as likely to run for Senate than are members in the
reference category. Indeed, Judiciary seems more closely related, in terms of its rela-
tionship to progressive ambition, to the exclusive and control committees than to the
rest of the committees in the House. Again, this may help explain Perkins’ (1981)
somewhat counterintuitive finding that, for all its seeming limitations, members of

Judiciary find it to be a choice assignment.

 

l7Rough because extreme policy positions need not necessarily be captured by the liberal-
conservative dimension which, by common practice, is held to be captured by the first dimension
NOMIN ATE score (Poole and Rosenthal 1997). Since many of the issues dealt with by the Judiciary
Committee touch on racial politics and social issues, I also estimated models that included a control
for extremity on the second dimension, which produced no change in the substantive findings.

18Estimating the model without these fixed effects produced no change in the substantive findings.

19As stated above, I exclude from the sample House members from states who have no Senate
election in a given cycle. I have also estimated the model with the full sample (N = 3249) but
including a dummy variable indicating whether the member’s state has a Senate election in that
cycle, which produced no change in the substantive findings.

20Further analysis using Beck, Katz, and Tucker’s (1998) recommended Binary TSCS estimation
technique produced results consistent with those in TABLE 1.

21Further analysis shows that the Budget committee’s inclusion in this category is driving most
of the effect.

22This is produced by calculating ef’iud‘mrv,

45

Table 2.1: Committee Membership & Higher Office Ambition’r

 

 

Logit Coefficient
[Robust St. Error]

 

Judiciary Committee 1.07***
[0.448]
Exclusive or Control 0.979***
Committee [0.262]
Committee Chair —1.12
[1.05]
Ideological Extremity —-1.97**
[0.864]
Open Senate Seat 2.18***
[0.322]
Senate Incumbent of 0.902***
the Opposite Party [0.286]
Freshman Senate Incumbent 0.333
[0.353]
State Presidential Vote -—0.005
of Member’s Party [0.017]
State Population (in thousands) —4.65E -— 5**
[2.11 E -5]
Constant —3.30***
[1.03]

 

N = 2159
Pseudo R2 = 0.195
Wald x2 = 173.5***
log pseudo-likelihood = -285.9
*1) < 0.10; **p < 0.05; ***p < 0.01 (one-tailed alternative)

 

1Notes: Logistic Regression. The data are pooled. The unit of analysis is the member. Robust standard errors
adjusted for clustering on the member are shown in brackets. Regional fixed effects 85 party x Congress fixed
effects are omitted from the table.

46

2.5 Substantiating Assumptions: Is the Judiciary
Committee Unique?

Despite the findings of the preceding section, it remains to be shown (rather than
asserted) that the House Judiciary Committee is unique in the mix of activities it
offers to its members. If such is not shown, then the Judiciary result of the last
section is open to criticisms about the idiosyncracies of the membership of Judiciary
and the general political context during the late-19808 and 1990s.

Having assumed that the mix of activities offered by Judiciary is unique, and
having shown that membership on the committee is associated with (at least in
my sample) a propensity to run for higher office, I now attempt to substantiate
this crucial assumption. One problem in doing this is that the activities Mayhew
(1974) describes are conceptual, rather than substantive, in nature. We have no
direct measures of position taking, advertising, and credit claiming. Instead, I offer
here some reasonable proxies of position taking and advertising, and assess whether
the Judiciary Committee in the House is significantly different from other House
committees on these measures.

First, in Table 2.2, I examine the number of hearings in House Committees from
the 80th through the 108th Congress. The data for this table are taken from the
Policy Agendas Project conducted by the Center for American Politics and Public
Policy at the University of Washington.23 The sample pools together hearings of the
full committee along with subcommittee hearings.

The intuition for examining the number of hearings as a proxy for the activities
associated with progressive ambition is that hearings provide a committee forum that
allows members to take positions and advertise themselves in a way that is similar to

speechmaking on the floor.24 Members who wish to engage more in these activities

 

23http://www.policyagendas.org/datasets/index.htm1\#congressHear
24An examination of floor speeches from the perspective of an ambition theory of legislative
organization seems an important extension of this dissertation.

47

should seek out committees that facilitate them and, at the same time, push the

committees they serve on to hold more hearings.

Table 2.2: Average Number of Committee Hearings per Congress, 80th-108th
Congress

 

 

 

Group Mean Standard Error N

Judiciary Committee 114.93 11.57 29

All Other Committees 67.56 2.55 599
t = 3.99

degrees of freedom 2 626
p-value < 0.0001

 

Table 2.2 shows the results of a simple two group t-test, comparing the average
number of hearings in the Judiciary Committee to the average in all other commit-
tees. Judiciary averages about 115 hearings per Congress, while all other committees
average about 68. The difference is hugely significant in the expected direction, which
lends some support to the uniqueness of the Judiciary Committee. For most of the
period of the modern Congress, Judiciary has held more hearings than other com-
mittees, potentially offering more opportunities for its members to engage in position
taking and advertising.

As an aside, the control committees (Appropriations, Way and Means, and Rules)
are not significantly different from all other committees. The Budget Committee in
the House actually has significantly fewer hearings each Congress than do other
committees.

The data for Table 2.3 are drawn from the Congressional Bills Project, hosted by

the University of Washington. Specifically, I present the number of bills reported by

48

House committees from the 80th through the 105th Congress.25 The sample includes
public and private bills, and covers the full set of Congresses currently available from
the Congressional Bills Project.

Here the focus is much more on position taking than on advertising. Committee
work on bills means taking committee votes and, perhaps more importantly, report-
ing bills paves the way for position taking opportunities on the floor. Reported
bills represent a stronger signal than, for instance, members’ introduced or cospon-
sored bills, which are relatively weak and costless signals. A committee that reports
many bills simply offers members more opportunities to associate themselves with
legislation that has passed many of the procedural roadblocks of the House. Since
reported bills are more likely to be considered on the floor, membership on a produc-
tive committee also means having more influence over the agenda and thus shaping
the position taking context of roll call voting on the floor. Again, the number of
reported bills is only a proxy measure, but as the results show, it further reinforces
the notion that the Judiciary Committee is unique in the mix of legislative activities
it offers its members.

Table 2.3 shows the results of a difference of means t-test, comparing the aver-
age number of reported bills in the Judiciary Committee to the average in all other
committees. Judiciary averages about 65 bills per Congress, while all other commit—
tees average about 17. The difference is hugely significant in the expected direction,
which again helps substantiate my assumption about the Judiciary Committee. In
the modern congressional era, Judiciary has reported more legislation than other
committees, meaning that members get more opportunities to take positions that

further their ambitions.

 

25http://www.congressionalbills.org/download.htm1

49

Table 2.3: Average Number of Bills Reported from Committee per Congress, 80th-
105th Congress

 

 

 

Group Mean Standard Error N

Judiciary Committee 65.15 4.72 26

All Other Committees 16.80 0.83 537
t = 12.33

degrees of freedom = 561
p—value < 0.0001

 

2.6 Conclusion

Committee membership is clearly related to the decision of House members to run
for Senate. My results provide support for the idea that members’ career tracks
influence their behavior within the institution (Hibbing 1986; Jenkins, Crespin, and
Carson 2005). The primary weakness of the analysis, though, is that the finding of
covariation is far from revealing the causal mechanism that relates particular com—
mittees to particular career paths. Even if we stipulate to the fact that certain of
Mayhew’s activities are more important for the pursuit of higher oflice, and that the
Judiciary Committee facilitates these activities, it is still difficult to distinguish be-
tween two causal stories. One states that some members begin with long term career
plans that include higher office, and so initially choose a committee like Judiciary
(or transfer to it at a later date) in order to improve their chances of achieving their
goal.

Another story could be told in which some idiosyncratic process, perhaps driven
by certain members’ interest in particular policy areas, leads members to join a

committee like Judiciary only to find, by accident, that the committee does facilitate

50

the pursuit of higher office, or even that these particular policy areas of interest lead
them to pursue higher office in order to wield more policy influence. A third story
could be told relying principally on the idea of opportunity costs. In this story,
Judiciary (or any other committee about which we might make a similar assessment)
is important not for the activities it facilitates but for the activity (namely, claiming
credit for distributive goods brought back to the district) that it decidedly does not
facilitate.

If we agree with Mayhew (1974) and Fiorina (1989) that credit claiming can
be decisive in winning reelection, then one’s current office becomes less valuable,
and spending time gaining seniority on Judiciary still does not improve one’s credit
claiming capacity. In that case, the member becomes more likely to run for higher
office because the office she is risking is less valuable than it would be if she was
serving on a committee more oriented toward distributive politics. In other words,
the opportunity costs are lower.26 Regardless, what these alternative arguments have
in common is that they all point to the idea that ambition theory, and progressive
ambition in particular, can provide insights into committee structure and committee
membership decisions.

While moving us in the right direction, the argument and results presented herein
are far from offering a full-fledged understanding of the role that ambition plays in
shaping the utilization of and, more importantly, the construction of the internal
institutions of legislatures. Judiciary is one of the oldest committees in the House of
Representatives, being established in 1813 (Nelson 1994). It is difficult to imagine
that the leaders of the early-19th century House intended for Judiciary to be an
‘incubator’ for future Senators. But institutional evolution and change often leave
us confused because of their often intangibility as phenomena. While its name has

remained constant, and its initial grant of jurisdiction remains, the Judiciary Com-

 

26While I have dealt with it only peripherally, this story could be connected to Rohde’s (1979)
idea of risk bearing in career choices.

51

mittee has evolved in terms of the issues it deals with and in how it deals with
them.27

Extensions of this analysis could improve the generalizability, or identify the po-
tential temporal specificity, of the results by expanding the timeframe of the data.
Certainly, drawing on Perkins (1981) for the background to the argument poses
that risk. Additionally, identifying the relationship between the House committee
on which they sit and the committee requests of progressive members once they
reach the Senate could demonstrate a linkage between informational theories (Gilli-
gan and Krehbiel 1990; Krehbiel 1991) and my theory of committees and progressive
ambition.28 The analysis could also be strengthened by bringing in data on House
members’ first run for political office (i.e. whether they first ran against a sitting
incumbent), data similar to those used by Rohde (1979) to measure members’ will-
ingness to bear risk, as well as data on committee requests. In both cases this
additional data will allow me to assess whether a higher level of progressive ambition
exists prior to taking a seat on a particular committee, or whether the committee
itself shapes the available strategies, and thus the decisions, of members of the House
of Representatives.

Most importantly, a better method of capturing the concept of ‘the progressive
ambition committee’ would strengthen future analyses. The prediction that the
Judiciary Committee is associated with a greater propensity to run for Senate is

plausible, and is borne out by the evidence, but such fails to capture the full range

 

27As Nelson (1994) points out, Judiciary gained jurisdiction over revision of the laws and immi-
gration and naturalization (in 1946) and over issues previously handled by HUAC / Internal Security
(1974). Whether concern for creating a ‘progressive ambition committee" entered into these deci-
sions is difficult to say, but it points to the problem of identifying distinct creative events in a
legislature. Indeed, an enterprise along these lines is made even more difficult by the fact that
much of jurisdiction is defined not by standing rules but by precedents set by the rulings of the
parliamentarian (King 1997).

28In other words, a member’s choice of Judiciary in the House is driven by their desire to,
eventually, gain a seat on the Senate Judiciary committee and thus play a role in shaping the
makeup of the federal courts, a role that is presumably very important to many members and their
constituents.

52

of considerations that could lead members to choose committees, even with the all-
consuming goal of moving up. A member from an urban district looking to run for
Senate could well engineer a transfer to the Agriculture committee, if doing so could
shore up support in the rural areas of the state. A member whose district mirrors
the whole state in terms of demographics, ideology, and economic interests should
lead the member preparing to run for statewide office to behave little differently than
other, statically ambitious members. A theory that would address this would have
to more explicitly deal with the conditionality of the choice to run for higher office,
and would require us to relate the committees to these conditions.

In addition, a quirk of the analysis suggests that membership on the House Budget
Committee is strongly related to the decision to run for Senate. This may be a
function of the policy activities of the Budget Committee, but there is reason to
suspect that it may also be, in part, a function of the peculiarities (especially in
the Democratic Caucus) of the process by which members are nominated to the
committee (Maltzman 1997). In addition, the limitation on how many terms a
member may serve on Budget, and the resulting rotation of members on and off
the committee, gives it a unique aspect. I suspect that the relationship between
membership on Budget and the pursuit of higher office relates to this committee’s
role as a ‘portfolio builder’. Members who serve on Budget can point to issues of
critical importance and say “I have worked on that”. But it should be noted that
this is distinct from Mayhew’s concept of credit claiming, which is of a decidedly
distributive nature. Portfolio building, in the sense in which I am using the term,
allows members to make credible claims to expertise in particular areas of policy
and about particular issues. But again, further analysis is required to explore this
phenomenon.

The results lead me to believe that congressional scholars should consider all the

goals posited by Fenno (1973), and additionally, should view the activities described

53

by Mayhew in the context of progressive ambition. The combination of factors that
influences the goals members pursue, and thus which committees are best suited to
facilitate the pursuit of these goals, needs to be more systematically addressed.

If Congress is organized to help members pursue their goals, and the pursuit
of higher office is a significant member goal, then institutions like the committee
and the party are likely to be structured in ways to facilitate this goal. But without
going back to first principles and considering that some legislators may be using their
institutional resources to pursue goals that we have long assumed away, we cannot
even speculate about the relationship between higher office goals and the design and

utilization of important legislative institutions.

54

Chapter 3

Progressive Ambition and Partisan
Theories of Legislative
Organization

3. 1 Introduction

In the 2004 presidential election, President Bush won 57% of the vote in the state
of Tennessee. In Tennessee’s 9th Congressional District, which covers much of the
Memphis metropolitan area, Bush won only 30% of the vote. The representative
from Tennessee’s Ninth District at the time was Harold Ford, Jr., a young Democrat
who seemed destined to move up in the political world. In 2006, an opportunity
presented itself for Ford, with Tennessee Senator and then Majority Leader Bill Hist
(R) choosing not to run for reelection. The open seat also drew Bob Corker, former
mayor of Chattanooga, into the race. For Ford it was a difficult race. His district was
significantly to the left of the state, at least in terms its preference for Democratic
presidential candidates, and yet Ford would continue to represent the 9th District
throughout his Senate campaign.

Ford’s campaign rhetoric, at least, reflected a move to the right in response to
this difference between his district’s preference and the preference of the statewide

constituency in Tennessee that would choose its senator.

55

Harold Ford Jr. ran hard to the right, talking incessantly about how
powerful “my Jesus” was, filming a campaign ad in a church, boasting
that he was pro-life, denouncing the New Jersey same-sex marriage ruling,

and wearing a camouflage hunting cap on Election Day.1

But what of his roll call behavior within the House? Is there evidence that Ford
matched this seeming change in rhetoric with a complementary change in how he
votes in the House? To illustrate some of the techniques I will use later in this
chapter, I will here present a simple analysis of Rep. Ford’s roll call behavior in two
consecutive Congresses, the 108th (when he was not running for statewide office)
and the 109th (when he was).

I use a scaling procedure called W—NOMINATE, which takes binary choice data
and unfolds it such that the choosers can be placed along one or more dimensions
according to their estimated ideal points (Poole and Rosenthal 1985, 1991, 1997).2
In typical practice, we focus on the ideal points that are estimated for the first
dimension, which is held to capture the general liberal-conservative orientation of
legislators. The procedure thus produces ‘scores’ for legislators (or whatever set of
choosers we have an interest in) that are bounded by -1 and +1, where left and right
are synonymous with liberal and conservative. To examine Rep. Ford’s preferences,
I use W-NOMINATE to scale first all roll calls for the entire 108th Congress, and
then (separately) all roll calls for the 109th Congress through September 30, 2006.
Comparing just the estimated first dimension ideal points, Ford moves from -0.422
in the 108th to -0.360 in the 109th Congress. But estimated ideal points of any kind
should be compared with some skepticism, and there is some debate among scholars

specifically as to the appropriateness of making simple over-time comparisons of

 

1Thomas F. Schaller, Salon, November 14, 2006, http://www.salon.com/opinion/feature/
2006/11/14/no_south/index.html

”The program for estimating W-NOMINATE scores, as well as the roll call matrices that are
scaled with the program and online manuals for implementing the procedure, are found at Professor
Poole’s website at http: //m.voteview.com/.

56

W—NOMINATE scores (Carson, Crespin, Jenkins, and Vander Wielen 2004; Herron
2004; Jenkins, Crespin, and Carson 2005; Poole and Rosenthal 1997; Rothenberg
2004; Rothenberg and Sanders 2000, 2004).

To go a bit further, then, I compare Ford’s estimated ideal points from the 108th
and 109th Congresses relative to the ideal points of two crucial actors-the median of
the whole chamber and the median of the membership of Ford’s party.3 In the 108th
Congress, Ford was 92% as far left of the median of the House of Representatives
as was the median of the Democratic Caucus. In the 109th, by contrast, Ford was
only 74% as far to the left. It is hard to doubt that this is the result of Ford moving
from representing a district that voted 30% for George W. Bush in 2004 to seeking
to represent a state that voted 57% for the Republican nominee. Ford went on to
lose on November 7, 2006, winning 48% of the vote to Corker’s 50.7%.

While Harold Ford’s run for Senate illustrates some of the ideas and techniques I
use in this essay, the analysis ignores one crucial element of my argument. It fails to
recognize that all roll calls are not the same, conflating different types of votes that
have very different functions both in the legislative process and in our theorizing
about it. Specifically, the simple calculation of Ford’s ideal points that I present
above includes procedural votes (among other types), on which voting considerations

seem quite different than those that drive voting on substantive matters (Froman and

Ripley 1965; Cox and McCubbins 2005).

3.2 The Danger of Sophisticated Legislating?

Surely in contemplating a sophisticated move in the legislative arena, a

legislator will give prior consideration to how it will play back home—and

 

31 measure this by calculating the following ratio for each Congress:

(Ford’sWNOM) — (Chamber’sWNOM)

.1
(Dem’sWNOM) — (Chamber’sWNOM) (3 )

 

57

to whether or not he can explain it satisfactorily to his supporters in
order to reduce the prospects that it will become a damaging campaign
issue. In short, the two-arenas hypothesis induces legislators to calculate

about being sophisticated (Denzau, Riker, and Shepsle 1985, 1118).

In Home Style, Richard Fenno (1978) presents a ‘two—arenas’ view of congres-
sional behavior, in which legislative behavior is viewed as instrumental in achieving
both legislative outcomes and reelection. For Denzau, Riker, and Shepsle (1985),
this raises the question of how constituents might view strategic behavior by their
representatives; whether this behavior will be viewed as instrumental in achieving
outcomes or as sincere position-taking. For Denzau et al., this question is explored
using amendment voting, addressing analytic results like McKelvey’s (1976) agenda-
controlling chairman.

But manipulating a series of binary votes is not the only way to use agenda
control strategically to achieve one’s preferred outcome. As Romer and Rosenthal
(1979) show, an agenda setter who can make a single ‘take-it-or-leave—it’ offer to
the median voter can achieve an outcome he prefers to the median voter’s ideal
point. The key to this sort of manipulation of outcomes, though, is the existence
of an institutional mechanism by which such an offer is possible. In the House of
Representatives, this mechanism is embodied in the special rule.

But the use of special rules to manipulate outcomes requires that members sup-
port their party’s efforts to do so. This same tension extends to weaker versions
of this argument, in which the decision not to grant a special rule works as a veto,
meaning that bias is achieved not in proposals (which end up where the median voter
wants them to) but in the selective targeting of status quopolicies to be changed
(Cox and McCubbins 2005). Either way, party members must acquiesce to the pro-
cedural maneuverings of their party in order for the party to bias the legislative

process.

58

As Denzau, Riker, and Shepsle (1985) might pointout, building on Fenno (1978),
there is a danger that cooperating with the party’s efforts to manipulate the process
may be taken as sincere position taking by one’s constituents (or that some other
knowledgable actor will try to convince one’s constituents to see it that way). This
danger exists for all members, but it seems especially likely to influence members’
behavior when they choose to leave their current office in pursuit of some other, higher
office. To the extent that party discipline over a member is exercised with implicit
or explicit threats to his position within the chamber (committee seat, committee
chair, access to floor consideration for his legislation, etc.), such discipline will not
work with a member who is leaving the chamber regardless. At the same time, the
member running for higher office has a new target constituency whose (presumably
more moderate) preferences he would like to account for in his legislative behavior.
And there are also likely to be interested observers who have an interest in pointing
out to that new constituency all the ways in which this ambitious member is wrong
for them.4

If members remain loyal to their party on procedural matters even when their vot-
ing on substantive matters shifts with the change in their new targeted constituency,
then it tells us something important about what motivates members to toe the party

line on procedure.

3.3 Previous Work

In Chapter 2, I argued that the literatures on strategic candidates and legislative
organization look past one another, particularly when it comes to the role played
by progressive ambition in shaping member behavior inside and outside the cham-

ber. Progressive ambition implies a member trading his or her commitment to one

 

4Senate elections are undeniably more competitive than House elections, with prominent, well-
funded incumbents and open seats that tend to draw quality candidates from both parties (Krasno
1994).

59

constituency for a commitment to a new constituency. This trade entails a probable
change5 in the constituency preference to which the member must respond in his or
her legislative behavior.

Consequently, research in the strategic candidates literature linking roll call be—
havior with progressive ambition predicts accommodation of the preferences of the
new constituency. But these analyses (Hibbing 1986) makes this prediction with-
out differentiating among vote types, and finds weak evidence for changes in behav-
ior, besides. Partisan theories of congressional organization (Rohde 1991; Cox and
McCubbins 1993, 2004; Aldrich and Rohde 2000) suggest that members’ behavior
should differ across different vote types without accounting for the desire of a strate-
gic candidate for Senate to adjust his or her legislative behavior to accommodate the
preferences of the new potential constituency.

Hibbing (1986) purports to identify the behavioral consequences of progressive
ambition for House members who make a run for Senate. He uncovers supportive
evidence of Schlesinger’s (1966) contention that the “ambitious politician must act
today in terms of the electorate he hopes to win tomorrow” (p. 6). The crucial
indicator used to identify this phenomenon is the ideological congruence of a House
member’s constituency with the statewide constituency. Behavioral change is ex-
pected to be greatest where the preferences of a House member’s constituency are
most different from the preferences of the state as a whole, and Hibbing’s evidence
bears out this expectation.6

Jenkins, Crespin, and Carson (2005) examine the roll call behavior of three differ-
ent groups of House members: those seeking reelection, those seeking higher office,
and those retiring. Their analysis focuses on how each of these groups respond to

procedural and substantive votes, and they find that Members who plan on retiring

 

5Indeed, likely change, usually in the moderate direction, due to the often profound difference
in size between district and statewide constituencies (Lee and Oppenheimer 1999).

6Hibbing measures ideology at the state and district level using the breakdown of the two-party
presidential vote.

60

at the end of a term are less likely to support their party on procedural votes than
Members from the other two groups. They also find, however, that Members seeking
higher office do not behave differently than Members seeking reelection to the House
when it comes to roll call votes, either procedural or substantive.

The contribution of this chapter is to provide the missing link between these two
representatives of the respective literatures as it relates to the effect of progressive
ambition on legislative behavior. Hibbing’s (1986) evidence for behavioral change
on the part of House members running for higher office is somewhat mixed until he
accounts for the congruence of district and state preferences, and Jenkins, Crespin,
and Carson (2005) find no difference between higher office seekers and reelection
seekers but do not account for this congruence.7

I seek to improve our understanding of the relationship between progressive am-
bition and partisan theories of legislative organization, then, by accounting both for
the difference between procedural and non-procedural votes as do Jenkins, Crespin,
and Carson (2005) and by also accounting for district—state congruence as does Hi—
bbing (1986). In terms of data, I rely on Poole and Rosenthal’s W—NOMINATE
procedure to calculate separate member scores for procedural and non-procedural

roll calls. I then use breakdowns of the two party presidential vote as a proxy for

district and state preferences.

3.4 Assumptions & Hypotheses

I rely on Cox and McCubbins’ (1993) party reputation argument as an important
causal mechanism that leads higher office seekers to maintain their support for
the party on procedural votes while changing their voting patterns to better align

themselves with the constituency of tomorrow, conditional on the congruence (or

 

7They do, however, find evidence for a difference between retiring House members and their
colleagues pursuing either reelection.

61

incongruence) of the preferences of their present district with those of their state as
a whole. I argue that party reputation is a resource from which a candidate benefits
regardless of whether he or she is running for House or Senate, so the expectation is
that on those votes most crucial to the party a member should act to preserve party
reputation (e.g. by voting with the party on procedural matters at the same level as
before he or she became a candidate for higher office). But on votes less crucial for
the party and its reputation, it is reasonable to expect a change in behavior on the
part of the candidate in the form of accommodation of the preferences of tomorrow’s
constituency.

I begin with a simple assumption supported by the literature on representation
in the American system (see Miller and Stokes 1963; Cnudde and McCrone 1966;
Weissberg 1978).

oAssumption 1: Current behavior by a Member of the House, whether
statically or progressively ambitious, already reflects accommodation of

that Member’s present constituency preference. ’

At the same time that members use their legislative behavior to accommodate
the preferences of their constituents, the party is also making demands on them.
In this way, partisan theories suggest that individual goals are not the only way
to motivate our understanding of legislative organization. Parties themselves help
determine organization of legislature, because members delegate to their party for
reelection purposes.8 Members are motivated to acquiesce to the demands of party
that are the product ofthis delegation relationship both because party reputation is
valuable to them and, more tangibly, because the party-in-the-legislature is tied to
the party-as-organization, which can provide additional electoral services to them.

Additionally, I rely on a two—part assumption drawn from partisan theories of

 

8Of course, what I will argue and subsequently test is the proposition that members also delegate
to the party to help them pursue higher office, not just for reelection purposes.

62

Figure 3.1: Partisan Theories & Change in Constituency Preference

 

‘ . BClosed. if) . 30126" is Si-Q
p i : i : i r

M F |2F-M|

legislative organization (Cox and McCubbins 1993, 2005) to further support the
proposition that the accommodation of a new constituency’s preferences is condi-

tional on the type of vote being considered.

oAssumption 2A: The majority party in the House of Representatives

controls the floor agenda through the use of special rules.

eAssumption 2B: The majority party in the House uses punishments and

rewards with the goal of achieving a unanimous coalition on rules votes.

The upshot is that party members remain free agents on some things, but not
on others. If the majority party in the legislature can use procedure to control the
agenda effectively, then it can often release members to vote with their constituency.
In Schlesinger’s terms, this could mean voting with today’s constituency or tomor-
row’s. But the majority party’s maneuvering may entail making choices on procedure
that are contrary to what a member’s target constituency would want, and thus cre-
ate tensions that may incentivize members to defect from the party. Figure 1 presents
a spatial illustration of this tension.

Figure 3.1 presents an example of how the decision to seek higher office should
influence the behavior of House members under this model, where M is the preference

of the majority party mediang, i0 is the present constituency (district) preference,

 

9Or of the majority party median’s district.

63

and i N is the new constituency (state) preference. When static ambition holds, i D and
i N are assumed to be identical, except in the case of redistricting (See Crespin 2005).
The interest here, however, is progressive ambition, where constituency preference
changes not because of redistricting but because the member has chosen to pursue
statewide office. The size and direction of i N — i D predicts the change in member
behavior. But under propositions 2A and 2B, the change in behavior posited for
final passage votes should not hold for special rules votes, regardless of the change
in constituency preference.

The example presented in Figure 1 depicts a policy dimension p where the ma-
jority party median (M) prefers moving the status quo SQ. If the majority party
brings a proposal to the floor under a closed rule (Bcjosed), the rule passes, and the
bill passes with the support of the floor median (F). If the majority party brings a
proposal to the floor under an open rule (Bopen), the bill is amended to F and passes.
iN prefers an open rule because IiN —- SQ] > IiN —- F] but [iN -— SQ] < [iN —— MI. in,
on the other hand, prefers a closed rule because lip -— M] < [in — F] < [in — SQI. If
the majority party can use punishments and rewards to hold together its coalition,
i N is compelled to vote for a rule that will produce a bill that is dispreferred to both
the status quo and, more importantly, to the bill that would be produced under an
open rule.

While I do not explicitly consider the case of minority party members here, I
think that much the same pressure is felt by these members. While their defection is
unlikely to change the ultimate outcome on a procedural vote, there may still be an
incentive to vote the party line, if we believe that party reputation (or at least the
portion of it that is influenced by the success and failure of the parties in the House)
is zero sum. Regardless, the best chance for the minority to thwart the majority’s
procedural tactics is to stick together. As attributed to Benjamin Franklin, “We

must all hang together, or assuredly we shall all hang separately.”

64

Now consider the above example not in terms of two different individuals, but of
the same individual whose ideal point has been shifted by his desire to win election in
a new constituency, thus shifting in the policy space from i D to i N. More specifically,
imagine this individual is a House member interested in winning a Senate seat with a
statewide constituency whose ideal policy is likely distinctly different from the ideal
policy of his House district.

Whether we imagine that special rules are principally negative agenda tools,
in that they selectively grant consideration to policy proposals on the basis of the
location of the status quo they target (Cox and McCubbins 2005), or as a positive
agenda tool that allows the majority party to make take-it—or-leave-it offers to floor
median (Monroe and Robinson 2006), we can see that whatever tension comes from
voting with the party on procedure is only exacerbated when a member runs for
office in a more moderate constituency (see Crespin 2005 for an application of this
argument to redistricting).10

Following this logic, alternative and null hypotheses can be constructed:

OHA: A new (potential) constituency preference induced by a run for
higher office produces static behavior on procedural votes but a change in
behavior on substantive votes consistent with the new constituency pref-

67‘67106.

0H9: A change in constituency preferences consistent with a run for
higher ofiice produces behavior on rules votes and final passage votes di-
rectly proportional to the change in constituency preferences, or no change

in other type of vote.11

 

10It seems clear, as Cox and McCubbins (2005) suggest, that there is some variation in the use
of special rules as either positive or negative agenda tools.

11Note that the effect of any selection bias (where the choice to seek higher office is endogenous
to the proximity of the policy preferences of one’s district to one’s state) is to create a harder test
for my theory: no change in behavior is necessary on any type of vote if i N — i D is already close to
zero.

65

The alternative hypothesis is really a test of a rather strong version of the party
reputation argument, using progressive ambition as a treatment. If members remain
loyal to their party on procedural matters even under conditions in which their
connection to the party-in—the—legislature has been severed and where they have
incentives to defect, then we have learned something we did not know about the key

role of party reputation in legislative decisionmaking.12

3.5 Analysis

The research design I utilize starts by assessing the change in roll call behavior,
and thus in estimated ideal points, from one Congress to the next.13 The design
is related to both the two-group pretest-posttest with untreated control group and
nonequivalent dependent variables quasi-experiments (Cook and Campbell 1979).
Some selection effects, particularly the potential relationship between a member’s
preexisting proximity to the statewide preference and his decision to run for Senate,
are handled by the pretest-posttest aspect of the design. This sort of selection effect
would mean that members who run for higher office would have to change their
behavior very little, and thus would bias against finding evidence in support of my
hypothesis.

The intervening treatment is a change in policy commitment to a new con—

stituency, with an expectation of how behavior should respond to the treatment.

 

12Interestingly, despite its critical importance in their partisan theories of legislative decision—
making, Cox and McCubbins (1993, 2005) have never subjected their party reputation argument to
a critical test. If party reputation helps members win reelection, and legislative success (however
defined) boosts party reputation, then there should be a direct relationship between legislative suc-
cess and the electoral success of members of the majority party. While my argument and analysis
might inform our understanding of party reputation, it is far from a critical test.

13This choice is made out of necessity. I would prefer to mirror the research design laid out
in Rothenberg and Sanders (2000), in which the last six months of consecutive Congresses are
compared to assess any change in behavior due to a member’s exit from the chamber (whether as
a result of retirement or running for higher office). There are too few roll calls within the subsets
I chose to limit the sample to the last six months of a Congress and still get the W-NOMINATE
routine to produce member scores.

66

Behavior is conceived of as a policy commitment (in the form of roll call voting) to
a constituency, and is measured with W-NOMIN ATE scores constructed separately
for final passage and rules roll calls.

Change in constituency is set to zero for members seeking reelection and is mea-
sured as the state presidential vote for the other party’s presidential candidate for
members seeking a Senate seat in the state. In this way, the intervening treatment is
analogous to variations in dosage in medical experiments. The control group, which
is made up of members of the House who do not run for Senate, receives a ‘dose’
of zero, while members of the treatment group receive higher or lower doses based
on the extent to which their state preferred the presidential candidate of the other
party in the last election.

I hold the commitment of members to their current constituency constant by
including the opposite party’s district presidential vote from the previous year. ’4

I look for evidence in support of the alternative hypothesis by running separate
seemingly unrelated regressions15 for Democrats and Republicans, where the depen-

dent variable is the change in the member’s W-NOMINATE score from one Congress

to the next, within the subsets of roll calls I identified.16 The data analyzed cover

 

14More specifically, for both statewide and district presidential vote, the measure takes on the
form of the two-party breakdown of the vote. So the vote for the opposite party’s presidential
candidate is represented as a proportion of the total votes cast for the presidential candidates of
the two major parties in the state or district:

PRESVOTERCP
(PRESVOTERCP + PRESVOTEDem)

 

(3.2)

for Democratic senate candidates and

PRESVOTEDem
(PRESVOTEReP + PRESVOTEDem)

 

(3.3)

for Republicans.

15I use Zellner’s (1962) method here because there is reason to believe that is correlation in the
errors of the final passage and rule votes equations. This is borne out by the correlation in the
errors yielded by the estimation, which was significantly different from zero.

1“There is justifiable concern about making cross-scale comparisons of W-NOMINATE scores
from Congress to Congress. But point estimate comparison (as opposed to distance comparisons
across scales) can arguably be done with fixed effects by Congress (or in this case, by congress-pairs)
and standard errors that are robust to heteroskedasticity (see Poole and Romer 1993; Poole and

67

the 10lst through the 107th Congresses, and the roll call vote types are identified
using data collected and provided by David Rohde (2005). The crucial relationship
is between Constituency Change and the dependent variable, AWNOMINATE. The
expectations are different for Democrats and Republicans. Accommodation of a more
moderate state constituency implies a rightward (positive) shift by Democrats and
a leftward (negative) shift by Republicans. The results of the analyses are presented
in Table 3.1. I control for the change in majority party after the 1994 election with
a dummy variable (Democratic Majority, =1 for the 10lst-103th Congresses). I also
control for the effect of state size and state heterogeneity (Lee and Oppenheimer
1999) by using the size of the state’s presidential electorate as a convenient proxy.
Included in the estimation, though not shown in Table 3.1, are fixed effects for each
pair of congresses for which the change in W-NOMINATE is calculated. In order to
establish the appropriate base category of members running for reelection to their
current district, I control for retiring members."

The models run on the Democrats in the sample provide supportive evidence
for the alternative hypothesis: Democrats who run for Senate move to the right
(move in a positive direction in the W-NOMINATE space) on their final passage
roll call behavior to better match their state, but do not do so in their rules voting
behavior. For Republicans there is a similar pattern in terms of the significance of the
coefficients, but the magnitude of the effects is very similar. As a Republican seeks
a Senate seat, and the state’s constituency is increasingly to the left, the member
is predicted to move leftward on substantive final passage votes, while no similar
change in behavior can be predicted based on these results.

My speculation about the difference in the magnitude of the effect of a Senate

run on roll call behavior between Republicans and Democrats yields two potential

 

Rosenthal 1997; Carson 2004; Jenkins, Crespin, and Carson 2005).
1"This mirrors the research design used by Rothenberg and Saunders (2000) and Jenkins, Crespin,
and Carson (2005).

68

Table 3.1: Preference Change in the House of Representatives

 

 

Seemingly Unrelated Regression]: Democrats, 1018t-107th Congress

DV: AWNOMINATE

Final Passage Votes

Special Rules Votes

Constituency Change 0.191"' 0.087
(State preference for Republicans I Senate run) [0.063] [0.067]
District Presidential Vote 0016* 0.109“l
(Republican) [0.040] [0.040]
Retirement -0.028* -0.012
[0.014] [0.014]

Size of Electorate -3.40E"6* 1.86E‘6

[1.78E-6] [1.86E‘6]

Democratic Majority 0.957"l -0.259"I
:1 for lOlst-103rd [0.020] [0.019]
Constant -0.421* -0.056
[0.026] [0.027]

N = 1562 N = 1562

R2 = 0.75 R2 = 0.33

 

 

Seemingly Unrelated Regressionj’: Republicans, 1018t-107th Congress

DV: AWNOMINATE

Final Passage Votes

Special Rules Votes

Constituency Change -0.071"' -0.074
(State preference for Democrats | Senate run) [0.042] [0.077]
District Presidential Vote —0.101* -0.082
(Democrat) [0.048] [0.087]
Retirement -0.004 -0.033
[0.012] [0.021]

Size of Electorate -9.38E‘7 -3.61E"6

[1.36E‘6] [2.491345]
Democratic Majority -0.034"‘ 0.378“
:1 for lOlst-103rd [0.020] [0.028]
Constant 0.032“ -0.405
[0.023] [0.045]

N = 1360 N = 1360

R2 = 0.29 R2 = 0.38

 

*Denotes significance at a = .05 (two-tailed) jNotes: The data are pooled. The unit of analysis is the member, with
each observation of the member being a pair of consecutive Congresses. Fixed effects for the Congress pairs are not

shown. Huber-White standard errors are shown in brackets.

69

explanations, the testing of which must wait for future work. First, I suspect that
the small state effect, meaning that House districts overlap with much of if not all
of the state constituency (as in at large House districts), is not fully controlled for
by my measure of state size, and that interaction effects may need to be taken into
consideration. If the small state effect (Fenno 1982) is different for Republicans
who consider a Senate bid than for Democrats, or if state size differently affects the
decision of Republican House members to seek Senate seats in the first place, then the
implications of a Senate run for roll call behavior among Republicans in the House
become difficult to assess. Second, and more intriguingly, if the statewide primary
constituency or the pool of potential intraparty rivals faced by the progressively
ambitious Republican House member is less tolerant of ideological apostasy in pursuit
of higher office than those faced by the typical Democratic House member, then

moderation may simply not be good strategy for Republicans.

3.6 Conclusion

Some important conclusions can be drawn from the findings presented here. First,
party reputation seems to be important for members seeking higher office. This
suggests that House members, at least, accept the idea that their party’s success in
the House influences their party’s reputation within the public, and that the benefits
of this reputation accrue whether they are running for reelection to the House or to
some other office.

Second, rules votes seem to be sufficiently hidden that they do not have the
impact they might in both of Fenno’s ‘arenas’. Even when we increase the potential
constituency pressure to defect from the party and at the same time increase the
danger of exposure by ambitious opponents or other interested actors, members still

remain loyal to the party.

70

Any study of how progressive ambition influences legislative behavior or legislative
organization faces the problem that many more members run for reelection at the
end of a given Congress, or even retire, than run for Senate or for any other higher
office. In the data I have compiled (10lst-108th Congress), only 90 members ran
for Senate. The question is whether enough members are motivated by progressive
ambition for us to ask that general theories take account of the desire of members of
legislatures to move up the career ladder. I argue that they do.

Despite useful theoretical frameworks that have treated them otherwise (see espe-
cially Mayhew 1974), legislatures are in reality mixed-goal institutions. Fenno (1973)
discussed five goals18 the pursuit of which shape legislative behavior, even in the US.
Congress. Schlesinger’s (1966) typology of ambition recognizes this mixed-goal dy-
namic. A very few discretely ambitious House members might seek to accomplish
one policy objective and then leave public office soon after. Most members of the
House appear to be statically ambitious, with varying levels of interest in policy ac-
complishment. Some set of members, not as numerous as the statically ambitious
but certainly more numerous than the discretely ambitious, looks upon the current
office as a stepping-stone for some higher office.

Working within the broader ambition theory framework provides a different way
of looking at how the institutional structures of a legislature are created and utilized
by members to pursue their goals. This is not to say that the goals and activities
associated with static ambition are of no importance for our theories of congressional
organization. I believe quite the opposite: that static ambition is in fact much
more important in explaining the contemporary Congress than either progressive or
discrete ambition is, thus Mayhew’s (1974) and even Fenno’s (1973) choice not to

use the higher office goal as part of their theoretical treatments of the House.

 

18Though often cited as discussing three goals, Fenno enumerates five: reelection, public policy,
power and prestige within the chamber, higher office, and personal gain; of which he only utilizes
three in his theoretical treatment of committees in the House.

71

Despite this, I hope that my continuing contribution will be to persuade that
it is an oversight to ignore the role that the insights of ambition theory can play
in our understanding of legislative organization more generally. Fundamentally, I
believe that by accounting for a fuller set of goals and ambitions, we can improve
our understanding of the way legislatures work, and that the value added is worth
the increased complexity of our subsequent theories.

In the end, I believe that congressional scholars need to consider the insight offered
by Aldrich (1993, 311), that “[T]hose...who have seen the ‘reelection imperative’ as
the major driving force for understanding legislatures (especially the United States
House) and the behavior of legislators in it can be seen as investigating a special case
of ambition theory.” Schlesinger’s ambition theory presents three different types of
ambition, one of which is central to our understanding of the modern Congress (static
ambition), one of which seems to be important at the margins (progressive ambition),
and one of which seems of very little importance to our understanding of the modern
Congress at all (discrete ambition). But the relative importance of these types of
ambition is, on some level, specific to the context of the modern Congress.19 By
broadening our conception of politicians’ motivations, we can move toward creating
a framework for the study of legislative politics that is applicable beyond the context

of the modern United States Congress.

 

19State legislatures in general, state legislatures with term limits (Herrick and Thomas 2005),
and the US. Congress at different points in history (Polsby 1968, 2004) could all be expected to
have different distributions of ambitions among their members than the modern Congress, and thus
different institutional needs/ demands among their members. A theory of legislative organization
that focuses on only one facet of ambition theory (i.e. static ambition) may thus be unable to
explain crucial aspects of these legislatures where other distributions of ambitions hold. Along
those lines, see Chapter 4.

72

Chapter 4

Toward an Ambition Theory of
Legislative Organization: How
State Legislative Term Limits

Show Us the Way Past the
‘Reelection Imperative’

[T]hose...who have seen the “reelection imperative” as the major driving
force for understanding legislatures ( especially the United States House )
and the behavior in it can be seen as investigating a special case of ambi-
tion theory (Aldrich 1995, 311).

4.1 Introduction

While some insights have so far been presented as to how a broader view of legislators’
motivations plays out in the context of the United States Congress, I have shown
nothing that suggests the need for a revolutionary rethinking of our understanding
of Congress. One might then question whether following Aldrich’s (1995) suggestion
that we view the reelection imperative as a special case of ambition theory adds
much to our understanding of legislative politics. The key, however, is that it is not
our understanding of Congress that needs a revolutionary rethinking, but rather our
understanding of legislatures generally.

More specifically, the theoretical toolkit that has been developed in modern con-

73

gressional scholarship must be augmented and improved before the insights of the
congressional literature can be brought to bear on the more general study of leg-
islatures. The reelection imperative has been a profoundly useful and successful
foundation for theories of congressional organization, but it is an assumption that
we know is false.1 I think that a broader view of legislative politics reveals an im-
portant property of the ambition theory of legislative organization. In particular, I
think that ambition theory can help build a theory of legislative organization that
can be generalized beyond a single, fixed set of institutions.

One way to demonstrate this theoretical advantage might be to apply an ambi-
tion theory of legislative organization to the institutional development of a legislature
over time. The problem with this approach, however, is that despite a long history
of institutionalization (Polsby 1968) punctuated by occasional and seemingly drastic
reforms (see Rohde 1991; Cox, and McCubbins 2005, Chapter 4), the fundamental
structure of the Congress today is arguably little changed from where it stood fol-
lowing passage of the 17th Amendment. This is because, while the typical style of
campaigning has certainly changed, and the relative balance of power within the two
chambers of the Congress among individual members of the House and Senate, the
committees, and the parties (majority and minority) have changed, the fundamen-
tal nature of what Schlesinger (1966, 5) calls “the electoral conditions which satisfy
[politician’s ambitions]” have not radically been altered.

In particular, Mayhew’s (1974) argument that members of Congress can be viewed
as single-minded pursuers of reelection was and remains fundamental to the study of

the Congress because the electoral conditions surrounding the pursuit and retention

 

1As a survey of the philosophy of science tells us, the simplifying assumptions of every theory are
false, on some level. In fact, theories are consciously false in their assumptions, in that they simplify
reality such that profoundly complex phenomena are rendered explicable, a analytic strategy that is
due to the limitations of human understanding. The most important consideration in constructing
new theories is whether assumptions that are marginally less false allow us to derive predictions
that are enough closer to empirical reality to justify the increased complexity of our theories. It
may be more art than science, however, to determine where increased predictive power justifies
increased theoretical complexity.

74

of a seat in Congress have changed very little, when viewed broadly and on a long
enough time horizon. Katz and Sale (1996) pinpoint the last monumental change
in the electoral conditions faced by members of Congress as the rapid adoption
of the secret ballot that took place around the country in the 18903, and use it
to demonstrate the organizational effects of the electoral connection. Indeed, the
consolidation of these effects, which can be observed in the existence of a significant
incumbency advantage, begin to be seen at least as far back as the 19208 (Cummings
1966).

I ask a question related to Adler’s (2002, 1) question, “Why is congressional
structure so ’sticky’?”, asking instead “When might legislative structure cease to
be sticky?” The answer I provide is that it is when legislators have no incentive
to maintain existing arrangements; when, in other words, the electoral connection
has been severed. Scholars have examined related questions often at the level of the
individual legislator, looking for evidence of ideological shirking by departing mem-
bers of the United States House of Representatives (Jenkins, Crespin, and Carson
2005; Rothenberg and Sanders 2000), but the wholesale dissolution of the electoral
connection that would be necessary to move from the erosion of representation of
constituents by particular legislators to a radical remaking of legislative institutions
is an extraordinarily remote possibility in the United States Congress. But such is
a reality in state legislatures that have faced the imposition of term limits, wherein
careerism as we have generally understood it becomes necessarily absent.

In this essay, I present what at first appears to be a novel test of the electoral
connection in the mold of Katz and Sale (1996). But the argument and the test of its
implications go further than that, to show explicitly what ambition theory can add to
our understanding of the connection between member’s motivations and legislative
organization. To state it another way, arguments about legislators’ utilization of

committees as a means to the achievement of their career goals need to account for

75

variation in those goals.

The argument is a simple one, and takes advantage of the natural experiment
represented by state legislative term limits. Despite its simplicity, though, the argu-
ment has a number of testable implications that I will discuss, only some of which will
be tested in the present essay. I begin by returning to the theme that animates the
larger project of which this essay is a part: questioning the fundamental motivational

assumption of modern legislative scholarship.

I shall conjure up a vision of United States congressmen as single-minded
seekers of reelection, see what kinds of activity that goal implies, and then
speculate about how congressmen so motivated are likely to go about
building and sustaining legislative institutions and making policy (May-

hew 1974, 5-6).

For Mayhew, this simplifying assumption implies particular activities in which
members of Congress engage in order to further their reelection prospects. These
activities in turn imply structural features of Congress as an institution that help

facilitate these activities.

4.2 An Ambition Theory of Institutional Mainte—
nance

As a political institution goes through a period of change, the functioning
of its subunits may be altered. Some of the institutional changes and re-
forms may be aimed directly at particular subunits. However, a subunit
may change because developments in the larger institution or other sub-
units cause or ofier it new roles to perform in the overall operation of the
institution (Oppenheimer 1977, 96).

Here, I present a theory of institutional maintenance that is intended both to
be general in its application but also to point to a particular set of predictions,

relating to legislative term limits, that diverge from those that would be derived

76

from the more traditional theories of congressional organization that have come out
of Mayhew’s (1974) electoral connection. In fact, if we take the electoral connection
at face value, any theory that takes it as its core assumption really cannot make
predictions about how a legislature should respond to term limits, institutionally.
Addressing this particular deficiency would be interesting, but I view term limits as
a way to address the more general question of how we should think about legislatures
in which reelection is not the only career goal.

Despite a move in recent years away from the “textbook” dominant committee
theories of legislative organization, it is still difficult to argue against the fundamental
importance of committees in legislative decisionmaking and in how members use the
institutional structure of the legislature to pursue their goals.2 While I have argued
that members of every legislature might have a variety of goals beyond simply that of
reelection, I have yet to make the crucial link between a more nuanced understanding
of legislators goals and the endogeneity of legislative institutions. If we understand
the structure of committees, in particular, to be the product of the set of legislators’
goals we assume, then changing our assumptions in this regard should change our
understanding of and our predictions about legislative committees.

Whether or not a particular structure (or aspect thereof) helps a legislator pursue
her goals is dependent upon what those goals are, and what sort of activities those
goals entail. As Mayhew has suggested, the goal(s) that we assume a legislator uses
her office to pursue determine the activities that our theory predicts she will engage
in.

A concern for the maintenance of a particular structure within a legislative insti-
tution is predicated on a concern for how that structure, as it currently exists, helps

a legislator pursue his goals. More specifically, a concern for individuals’ committee

 

2Indeed, even partisan theories of legislative organization recognize the importance of committees
and committee seats to members, arguing that providing or taking away a committee seat is an
important tool in enforcing party discipline. (Rohde 1991; Cox and McCubbins 1993, 2005).

77

seat property rights is largely driven by careerism (Katz and Sale 1996), or, in other
words, the electoral connection (Mayhew 1974). At the same time, a concern for
the maintenance of committees’ jurisdictional property rights is predicated on this
concern for individuals’ committee seat property rights.

So careerism (or the electoral connection) affects both of the sorts of committee
property rights we are interested in. “Careerism” and the “electoral connection”
are synonymous with static ambition (Aldrich 1995, 311; Schlesinger 1966). But
Schlesinger ( 1966) points to two other sorts of ambition that politicians might have
instead of static ambition—progressive ambition, which entails the desire for higher
office, and discrete ambition, which entails politicians viewing their service in their
current office, and their ‘career’ in politics as a temporary thing.

We could imagine the membership of any particular legislature as distributed into
these three different categories of ambition, and think about how that distribution
will produce very different institutional demands. If we believe the arguments of
congressional organization scholars, then the prevalence of static ambition should be
directly proportional to a concern for property rights of both sorts. The reason why
the reelection imperative fits the contemporary US. Congress is not that it charac-
terizes every legislature well, but that assuming that every member is a single-minded
seeker of reelection approximates the distribution of members into the categories of
discrete, static, and progressively ambitious. But we know that at least some legisla-
tures are characterized by a different distribution of ambitions. In particular, there
are 15 legislatures in the American states in which term limits have imposed on leg-
islators. In these legislatures, then, the distribution of ambitions has been artificially
changed. There can be no static ambition, no careerism, in a legislature in which
members only serve 6 or 8 years.

As static ambition is removed by term limits, then, changes that undermine

property rights will become more common. More specifically for my current focus,

78

implicit queues for membership will become expanded committees. At the same time,
when the electoral connection is severed there will be more pressure to create new
committees, which have the effect of slicing more thinly the fixed pie of legislative
jurisdiction.3

In a careerist body, what the individual member should want is to maintain their
influence over (and ability to claim credit for) the spoils that can be brought back
to their constituents and, more generally, the policy that is made in a particular
issue area. So keeping a seat on a particular committee could be less important
to a legislator in a context where the committee system is already volatile. What
a member should work to prevent, however, is the dilution of his influence and
credit claiming. Munger (1988), following Shepsle (1978), refers to this as currency
inflation, and argues that members should exercise “counterexpansionist pressure,”
to preserve the entitlements that make a committee seat valuable to members.4

Munger states the argument explicitly, writing, “The larger the increase in seats
on a committee, the more an individual assignment to that committee declines in
value, all else being equal” (1988, 337). But this argument is premised on some crucial
assumptions that remain implicit, the most important being that the dominant goal
among members is to be reelected to their current office.

If we believe that a concern for individual committee seat property rights is a
necessary condition for a concern for committees’ jurisdictional property rights, then

what happens when we change our assumptions about members’ goals? How should

 

3This ‘fixed pie’ assumption is very important, but I think it is justifiable. Legislatures do not
often invent new policy areas on which to legislate, though post-2001 some state legislatures did
create committees dealing with homeland security, which may have represented an expansion of
the total jurisdiction of the legislature. Even if we recognize that jurisdictions are not as sharply
delineated in some legislatures as they are in the US. House (Squire and Ham 2006), there is still
a sense in which consideration of a bill by one committee means another committee cannot consider
it, thus threatening the ability of some members to claim credit. And creation of a new committee
must mean that this dynamic becomes more intense.

4Speaker Reed once referred, in discussing the committees in the House, to “the very great
inconvenience of numbers.” (quoted in MacNeil 1963, 149). We might interpret this as both a
statement as to the effectiveness and efficiency of committees acting as ‘little legislatures’ and as a
characterization of the interests of committee members.

79

committee jurisdictions be affected? Maintaining the responsibility for particular
issues within a particular committee seems to be of greatest importance for those
who have paid their time to gain seniority within the committee (i.e. chairs, vice
chairs, ranking members). But paying in one’s time becomes irrelevant in the context
of term limits, since whatever power accrues to one as chair or ranking member will
evaporate once term limits come into play. I

Instead, members should come to view committees as fora in which to advertise
and take public positions, which are the activities Mayhew (1974) associates with
higher office ambition. The exclusivity of committee jurisdictions become less im-
portant, so long as progressively ambitious members are given the opportunity to
serve on the committees that facilitate the positions they want take. Even discretely
ambitious members, or at least those among them who have some policy agenda, will
be satisfied as long they get to serve on the committee that has jurisdiction over the
relevant policy.

A more general theory like the one I am advancing allows us to specify conditions
in which the effects that Shepsle, Munger, and Katz and Sala posit are attenuated by
the presence of alternative goals among members of a legislature. Even in a careerist
legislature, not every member views serving in their current office as a career. Some
view it is a public service with, presumably, a relatively brief time horizon. Other
members view serving in their current office as a means to winning some other office.
To the extent that this variation in goals is even a possibility in some legislatures,
it suggests that we need a more general theory than the one that has been built to
explain the US. Congress and assumes the reelection imperative. If we believe that
this assumption should be broadened, then it suggests some ways in which we need
to rethink our Mayhewian understanding of legislative organization.

Ambition theory provides us with a ready-made framework for a more general

theory. It retains one of the most useful properties of the reelection imperative, in

80

that it views the full set of activities associated with legislating as instrumental in
achieving career goals. Term limits provide us with a natural quasiexperiment to
test the predictions of an ambition theory, in which careerism / the reelection impera-
tive / static ambition is effectively eliminated as a goal. If ambition theory can inform
our understanding of legislative organization, then, term limited state legislatures

are the place to find out.

4.3 Hypotheses & Data Analysis

If the concern of an individual legislator for his committee seat property rights is
a necessary condition for a committee as an entity to have a concern for its juris-
dictional property rights, and careerism as Katz and Sala (1996) use the term is a
necessary condition for an individual legislator to have a concern for his commit-
tee seat property rights, then both individual and committee property rights should
become (at minimum) less important as a consideration when careerism disappears.
Term limits eliminate careerism in a legislature. Members are forced, in Schlesinger’s
(1966) terminology, to become either discretely or progressively ambitious when term
limits are imposed upon them. Members must either use their current office as a
springboard for some other office, or must view it as a short-term position which
they use to fulfil their civic duty or to pursue some limited set of concrete policy
goals. If committee seats are used for purposes other than the pursuit of a career in
one’s current office, then members should care less about protecting both the ‘value’5
of their own committee seats and the value of their committees’ jurisdictions.“

Following this logic, two different alternative hypotheses can be constructed.

 

5In the sense that both Shepsle (1978) and Munger (1988) suggest the use of the term ‘value’.

6Why would the policy-oriented discretely ambitious member acquiesce to the erosion of com—
mittee property rights with respect to jurisdiction? All the individual member should care about is
membership on the committee that at any given time has jurisdiction over the policy area in which
he has policy goals. Whether that committee had the jurisdiction previous to the consideration of
his policy or whether it has it after its enactment is irrelevant.

81

oHypothesis 1: When careerism as a possible motivation for legislators is

eliminated, then the size of committees should increase, all else equal.
and

oHypothesis 2: When careerism as a possible motivation for legislators is
eliminated, then the creation of new committees should be more common,

all else equal.

For Katz and Sala (1996), the reason careerism leads to a concern for committee
property rights is that committee members use their positions to claim credit for pro-
viding something of value to their constituents.7 Working in that context, Mayhew

(1974) writes:

Whatever else it might be, the quest for specialization [in a legislature]
is a quest for credit. Every member can aspire to occupy a part of at
least one piece of policy turf small enough so that he can claim personal
responsibility for some of the things that happen in it. Better yet, he
can aspire to rise in seniority and claim even more responsibility—perhaps

even be christened a “czar” or a “baron” by the press (95).

Dilution of credit in that context can be a dangerous thing, as Mayhew’s charac-
terization of the quest for specialization implies. Whether it is claiming credit for
district projects or exercising influence over a bureaucracy over which the committee
has oversight, more demands from within a particular committee can push up against
the limits of supply.

But when reelection as a career strategy is limited, members care less about

what their current constituents want and start thinking, to paraphrase Schlesinger

 

7Recall that their argument relates to the increased ‘personal vote’ nature of congressional
elections after the introduction of the secret ballot, creating a closer representative relationship
between members and constituents.

82

(1966), about tomorrow’s constituents. Perhaps that means a focus on a statewide
constituency, or a focus on the primary constituency in some other office, but it
certainly means that the legislator’s focus is not so narrow as we would expect it
to be if a career in her current office was the priority (Carey, Moncrief, Niemi, and
Powell 2003).8 At the same, if tomorrow’s constituents don’t exist, as in the case of
a member who is leaving political office and public life, then members presumably

care even less about what today’s constituents want (Rothenberg and Sanders 2000).

4.3.1 Individual Committee Seat Property Rights

In order to test the first hypothesis, I gather data on committee membership from
the Michigan House and Senate from 1993 to 2007.9. The Michigan legislature is
considered one of the most professionalized state legislatures in the US, in terms of
members’ salary, staff, and the time spent in session.10 While an analysis of a highly
professionalized legislature might threaten the generalizability of the inferences to
other, less professionalized state legislatures, it may allow us to approximate a coun-
terfactual in which another, highly professionalized legislature that has evoked much
interest in legislative studies, the United States Congress, is imagined to have term
limits imposed upon it.

The research design approximates an interrupted time series, in which a single
group is given repeated pretest measurements, is administered a treatment, and then
again given repeated posttests. The treatment is the impact of term limits, which
in Michigan first occurs in the 1998 election cycle for the House, and in 2002 for the

Senate.11 The unit of analysis is the committee, and the dependent variable through-

 

8Herron and Shotts (2006) present a model which suggests that term limits, which limit ca-
reerism, can produce certain conditions under which the provision of district-targeted pork is re-
duced.

9At this writing, data from 2003 are not available, so I have had to leave the committees from
2003—4 out of the analysis

10See the National Council of State Legislatures, “Full- and Part-Time Legislatures, ”http:
//www.ncsl.org/programs/press/2004/backgrounder_fu1landpart.htm

11So the impact on committee membership should occur with the organization of the chambers

83

out this subsection is the number of members on the committee at the beginning of
a legislative term.

Before the impact of term limits, the average committee in the Michigan leg-
islature had 9.08 members, while after the average committee had 10.81 members.
Looking at just the bivariate relationship, then, between term limits and the size of
committee rosters using a difference of means t-test with a directional hypothesis,
there is a significant difference after term limits: tdf:331 = 2.86, Pr(T > t) = 0.002.

Because we have neither random assignment nor a control group, the econometric
method and especially the control variables in the model become important. In
the bivariate relationship, we cannot distinguish between House committees and
Senate committees, nor can we single out exceptional committees like Appropriations
(always larger than other committees) or “housekeeping” committees like the House
Oversight & Ethics Committee (almost always smaller than other committees). At
the same time we cannot account for potential nonindependence within legislative
terms, nor for any factor that might reveal the spuriousness of the hypothesized
relationship.

Because the measure of the dependent variable is a discrete count, there are legit-
imate doubts that ordinary least squares regression is the appropriate quantitative
technique. With count data, then, there are two generic techniques that could be
appropriate—Poisson regression and negative binomial regression. In choosing be-
tween them, the distributional assumptions of the Poisson model are crucial. To use
the Poisson model, the count data in which we are interested should be distributed
such that the mean and variance of the dependent variable are equal. Looking at the
data, the mean number of committee members in the full sample is 9.9, while the

variance is 31.3.12 In the presence of overdispersion like that present in these data,

 

in 1999 and 2003, respectively.

12Interestingly, the mean for just the Senate sample is 6.1 with variance 6.5. This is probably
close enough that the Poisson model would be appropriate. For the Michigan House, though, the
mean is 12.8 with variance 30.5.

84

the alternative to the Poisson model is negative binomial regression (NBR), which
weakens the distributional assumption of the Poisson.

An additional quirk of the data that requires attention is the absence of zeros in
the dependent variable measure. A committee with zero members ceases to exist,
and thus is not present in the data. But NBR can produce predicted values of
zero, which makes little sense in this context. An alternative is to use a technique
called zero—truncated negative binomial regression (ZTNBR), which accounts for this
truncation of zeros in count data.13

Equations 1 and 2 present the specifications of the empirical models that I esti-

mate via ZTN BR.

31 = a+filX1 +B2X2+53X3+54X4+B5X5+e (4.1)

and

y = a + IBIXI + )62X2 ‘l‘ ,83X3 + @1ij + ,65X5 + ,BfiXfi + [67X7 + E (4.2)

In Equation 1, X1 is an indicator of the impact of term limits, X2 indicates
whether the committee is in the Senate, X3 indicates whether the committee is the
Appropriations Committee in either chamber, X4 whether it is a “housekeeping”
committee, and X5 measures the number of committees in the chamber in a given
legislative term.14 Equation 2 takes this specification and adds some partisan con-
trols. X6 measures the percentage of the seats in the chamber held by the majority
party, and X 7 indicates a Republican majority party in the chamber. The empirical

results for these two models are presented in Table 4.1.

 

13See Long and Freese (2006) for a discussion of the technique and its implementation in STATA
9.

14Controlling for the number of committees is crucial, as changes in the size of committees could
be produced by an expansion or contraction of the number of committees, since there is a fixed
number of legislators. Not accounting for this could threaten an inference we might make about
how the dilution of the value of committee seats is being driven by ambition considerations.

85

Table 4.1: Committee Size in the Michigan Legislature, 1993-2007

 

 

Zero-truncated Negative Binomial Regressions’f, DV: # of Members

 

Expectation Coefficient Coefficient
(Robust SE) (Robust SE)
Term Limits + 0.069** 0.063**
(0.018) (0.021)
Senate — -0.684** -0.727**
(0.024) (0.035)
ApprOpriations Committee + 0.897** 0.897**
(0.028) (0.028)
Housekeeping Committee — -0.478** —0.478**
(0.145) (0.145)
Number of Committees +/— 0.015** 0.007
(0.002) (0.004)
Majority Party’s % of Seats +/— 0.005
(0.003)
Republican Majority +/— -0.063*
(0.026)
Constant +/— 2.11** 2.165**
(0.054) (0.194)
Observations 333 333
Log likelihood -826.7 —826.3

Robust standard errors in parentheses (adjusted for clustering on 16 chamber-terms)
* significant at 5%; ** significant at 1% (two-tailed alternative)

 

fNotes: The data are pooled. The unit of analysis is the legislative committee (House & Senate).

86

Before discussing the results, one more thing needs to be pointed out. The robust
standard errors shown in parentheses are adjusted for clustering on the chamber-
term. Clustered standard errors are often used in survey research to account for cor-
relation or nonindependence of observations within some subset of the full sample—for
instance, respondents from the same family. The technique, first explicated in the
finance literature (Froot 1989) is more generally applicable wherever we suspect non-
independence within subsets of our observations (Williams 2000; Wooldridge 2002).
Here, there is reason to suspect that while committee rosters are not zero-sum (in
fact, the theory explicitly predicts them not to be) they are nonindependent in that
putting a member on one committee makes it marginally more difficult for that mem-
ber to serve on another committee given, if nothing else, the constraints of time.

Moving to the results, then, I find support for the individual committee seat
hypothesis. In both models the coefficient associated with the term limits indicator
is positive and strongly significant at the 1% level (at least). The control variables
for which I have expectations produce coefficients that work how I expect, in that
Senate committees are predicted to be smaller, the Appropriations Committees in
both chambers are expected to be larger and housekeeping committees smaller. The
partisan control variables do not undermine the term limits result, though the Re-
publican Majority variable produces an intriguing result, in that committee rosters
are predicted to be smaller under Republicans.

Overall, Hypothesis 1 is strongly supported by these models.

4.3.2 Committee Jurisdictional Property Rights

Hypothesis 2 is somewhat harder to test empirically, as committee jurisdictions in
the states are harder to assess than they are in the US. House or even Senate (Squire
and Hamm 2005). Instead of looking for jurisdictions in rules or precedents in the

legislatures I examine, I make an assumption that the total jurisdiction of a legislature

87

is a fixed pie, and that every committee has a slice of a given size. If this assumption
holds, then the creation of a new committee means that some other committee has
lost some of its jurisdiction. So to test Hypothesis 2, I examine committee creations
in 4 state legislatures.15

The data are drawn from both chambers in California, Maine, and Ohio from
1993 through 2005, and in Michigan from 1993-2007. I count committee creations
each term by looking at either the legislative manual published at the beginning of
each term or at the rules adopted at the start of the term and published in the floor
record.16

Before the impact of term limits, the average number of committees created in a
chamber at the beginning of a legislative term was .82, while after the average was
1.33 committees. The difference of means t-test of the relationship between term
limits and the creation of new committees yields tdf=64 = 1.24, Pr(T > t) = 0.110.
While this is close to statistical significance, it is not nearly so clear a relationship
as that found above, between term limits and the size of committee rosters.

Again, these are count data, which points toward either a Poisson or a negative
binomial model. Looking at committee creations, the mean in the sample is 1.08,
while the variance is 2.88.17 Since there is some evidence of overdispersion, I use the
negative binomial model.

With these data there is also an issue with zeros, but this time there may be too

many. Over half of the observations in the sample have a zero count for committee

 

15This throws out, then, jurisdictional changes in which jurisdiction over some set of policies is
taken from one committee and given to another. Changes of this sort probably reflect the balance
of power within the chamber more than the diminished importance of committee jurisdictions,
generally.

16I also counted the creation of Joint Committees in Maine, since they play an important role in
the legislative process (Squire and Hamm 2005).

17Jumping ahead slightly, when the full negative binomial regression model presented in Table 2
is run, the test of a, which is the overdispersion parameter for the negative binomial model, yields
a test statistic 2%,) = 3.075, p-value = .04. While significant, it is rather marginal for these sorts 0
tests. In fact, running the Poisson model provides stronger support for Hypothesis 2 than negative
binomial regression, but I present the more conservative test here.

88

creations. There is another variant of the negative binomial model that can account
for this potential problem, which could lead to the underprediction of zeros by the
model. This alternative is called zero-inflated negative binomial regression (ZINB).
But running this model and conducting the Vuong test outlined by Greene (1994)
suggests that ZINB is not an improvement on the standard model (2 = 1.16, p—value
= 0.124).18 I therefore estimate negative binomial models on the pooled observations

with the following specifications.
9 = a + £3in + fizXz + [3ng + 54X, + 05X, + HGXS + [37X7 + figxg + e (4.3)

and

y = a +fi1X1+fi2X2 +B3X3 +,64X4 +fi5X5 +fi6X6 +,87X7 +58X8 +BQXQ +6 (4.4)

In Equation 3, X1 is an indicator of the impact of term limits, X2 indicates whether
the chamber is the upper chamber in the legislature, X3 indicates whether majority
control of the chamber has switched parties since the last term, X4 measures the
percentage of the seats in the chamber held by the majority party, X5 indicates a
Republican majority party in the chamber, while X6, X7, and X3 are dummy vari-
ables for California, Maine, and Michigan (leaving Ohio as the excluded category).
Equation 4 includes an additional control for X9, which is the number of committees
eliminated in a chamber at the beginning of the term. The estimates are presented
in Table 4.2.

Hypothesis 2 does not fare so well as did Hypothesis 1. The first model shows
a significant relationship between the effect of term limits and committee creations
(one-tailed p—value of 0.038), but it is near the margin of significance. But remember

that we are assuming a fixed pie of jurisdiction. Because of this, I must control for

 

18I ran the ZINB model’s inflation equation with a constant only. This model also produced

stronger evidence of the effect of term limits than the models presented in Table 2.

89

Table 4.2: Creation of New Committees in the California, Maine, Michigan, & Ohio

Legislatures, 1991-2007

 

 

Negative Binomial Regressionf, DV: # of New Committees

 

Expectation Coefficient Coefficient
(Robust SE) (Robust SE)

Term Limits + 0.530* 0.463
(0.298) (0.294)

Senate +/— -0.938** -0.761*
(0.371) (0.351)

Postswitch + 0.612* 0.358
(0.352) (0.419)

Majority Party’s % of Seats +/ - 0.065 0.064
(0.063) (0.061)

Republican Majority +/ —- 0.212 —0.222
(0.488) (0.530)

California +/ — -0.219 —0.532
(0.667) (0.666)

Maine +/— —0.209 -0.110
(0.613) (0.572)

Michigan +/— 1.32* 1.011
(0.562) (0.632)

# of Committees Eliminated + 0.141*
(0.073)

Constant +/— -4.37 4.03
(3.80) (3.64)
Observations 66 66
Log pseudolikelihood -81.9 -80.0

Robust standard errors in parentheses

* significant at 5%; ** significant at 1%
(Term Limits: one-tailed alternative, all others: two-tailed)

jNotes: The data are pooled. The unit of analysis is the legislative term for each

legislative chamber (House/ Assembly & Senate) within the timeframe.

90

the number of committees that are eliminated from the chamber in a given term,
else the assumption is not reflected in the empirical analysis. When I control for
the elimination of committees, the effect of term limits moves to the other side of
marginal statistical significance (one-tailed p-value of 0.058).

Why might this be? Perhaps there is still reason for maintaining committee
jurisdictions after term limits that is unaccounted for in my theory. Perhaps the
number of observations here is too low to obtain reliable estimates.” It is safest to

say that the test of Hypothesis 2 provides suggestive evidence in support.

4.4 Discussion

The argument and the analysis presented here essentially reverse Katz and Sala’s
(1996) argument about committee property rights. For them, the introduction of
the secret ballot in the United States at the end of the 19th century created in-
centives for members of Congress to protect their committee property rights. This
was because the secret ballot eroded the almost deterministic relationship between a
party’s electoral success and its members’ success that was created by party—printed
ballots (Cain, Ferejohn, and F iorina 1987). With the rise of the personal vote and
candidate-centered elections, maintaining a long-standing relationship between rep-
resentatives and constituents became imperative. One important way of doing this
was for members to find a means of exercising influence over policies of importance to
their constituents, and to make sure that some of the credit for policy achievements
came to themselves. Committees provided this means, and members had an incen-
tive to ensure that they received a seat, kept a seat, and maintained the value of a

seat on the committee or committees that were most important to their constituents.

 

191 am in the process of collecting data from other states, particularly New Hampshire and
Pennsylvania, which do not have state legislative term limits. More states, and particularly those
without term limits, would help mirror stronger quasiexperimental research designs by providing a
control group.

91

With the imposition of term limits, the process unravels, but not because the
personal vote is no longer important. What happens, instead, is that members must
pursue personal votes from a wider pool of voters. In this context, the value of a
committee seat is no longer figured in terms of seniority or credit claiming, since
both are associated with the needs of a member who hopes to make a career out
of serving in her current oflice. Whether it is position taking, advertising, portfolio
building, etc., the activities in which term limited members need to engage are not
aided by maintaining committee property rights, either individual or collective. For
state legislative committees after term limits, it comes down to story of ‘the more,
the merrier.’

I began with the idea that the reelection imperative is a special case of ambition
theory (Aldrich 1995). Given the position of the reelection imperative in theories of
legislative organization, this idea has important consequences. Beyond our positive
theories of politics, how would an ambition theory of legislative organization inform
normative theories of representation? Legislators as delegates would seem to be
much less common when forms of ambition other than static ambition are prevalent
in a legislature. 'Ifusteeship may characterize the relationship between at least the
discretely ambitious politician and her constituents. But on many matters, shirking
may better characterize the constituent-representative relationship for members who
are not statically ambitious. Indeed, given the evident erosion of committee-related
property rights, there may be reason to expect that the dilution of the value of an
individual committee seat is reflected in a reduction of the influence of constituents
over policy areas in which they have very real interests at stake. As Carey, Moncrief,
Niemi, and Powell (2003) and Herrick and Moore (2005) show, term limited legisla-
tors espouse different attitudes about representation and demonstrate different sorts
of ambition, but there is justifiable doubt about whether term limits advocates have

gotten what they sought by imposing term limits.

92

Chapter 5

Conclusion

5. 1 Summary

In this dissertation, I have presented an ambition theory of legislative organization. I
began with the premise that politicians use the institutions around them, and in fact
sometimes change the institutions, in order to pursue their goals. I challenged the as-
sumption in contemporary congressional scholarship that members are single-minded
seekers of reelection, arguing instead that the broader conception of politician’s ca-
reer goals offered by Schlesinger’s (1966) ambition theory could provide a foundation
for legislative theories that are more broadly generalizable.

The substantive portion of the dissertation (Chapters 2-4) was divided into three
essays. Chapter 2 presented two pieces of evidence supporting the contention that
members of Congress use committees to pursue goals other than reelection, and that
particular committees that facilitate the goal of higher office, in particular, can be
identified. First, I presented a case study of Richard Nixon’s membership on the
House Committee on Un—American Activities (HUAC), showing that the committee
did seem to facilitate position taking and advertising, the two activities Mayhew

(1974) pointed to as most necessary in using one’s House seat as the means to win

93

higher office. Nixon, himself, and many other observers made points that relate to
the argument that particular institutional structures are used, and may be designed,
to facilitate these activities.

In the same chapter, I presented a quantitative analysis that showed evidence in
support of the prediction that another committee, the House Judiciary Committee,
similarly facilitates the pursuit of higher office. Judiciary is a relatively high demand
committee, despite the lack of credit claiming opportunities for its members. Anec-
dotally, at least, the ‘style’ of Judiciary Committee politics seems similar to that
of the now-defunct HUAC. In fact, the jurisdiction of HUAC was absorbed into the
Judiciary Committee’s jurisdiction during the mid-19703 reforms of the House, which
is suggestive.

In Chapter 3, I presented an argument that synthesized my recurring ambition
theory argument with partisan theories of legislative organization. Specifically, I ar-
gued that the maintenance of party reputation, which creates incentives for members
of the House to delegate institutional power to their party leaders, remains valuable
to members who choose to run for higher office. This is despite the severing of the
formal relationship of that member to their party caucus and their party leaders,
and despite the change in constituency preference that their pursuit of a Senate seat
often entails. I tested a prediction derived from this argument, that progressively
ambitious members will adjust their substantive roll call behavior in accordance with
the preferences of the statewide constituency while remaining faithful to their party
on crucial procedural votes. The prediction was supported by the empirical evidence.

In Chapter 4, I used the outlines of an ambition theory of legislative organization
to address the endogeneity of legislative institutions. Focusing on the maintenance
of both individual and collective committee property rights, I argued that term
limits tend to create pressure to ‘devalue’ committee seats and committee jurisdic-

tions. Term limits artificially structure the distribution of ambitions in a legislature,

94

transforming even the most professionalized, careerist legislative body. Members are
forced to become either discretely or progressively ambitious, and they will push
to redesign their institutional environment in ways that facilitate these ambitions.
Since members feel less need to engage in the sort of credit claiming activities that
make seniority and smaller committees necessary, committee membership instead
becomes a means of engaging in position taking, advertising, and portfolio building
(or in the case of the discretely ambitious legislator, to achieve specific policy goals).
I tested predictions about the increasing size of committee rosters and the creation
of new committees in four term-limited legislatures, finding suggestive evidence that
committee creations become more common after term limits and strong evidence

that committee rosters increase in size after term limits.

5.2 What I Have Learned

When I first conceived this project, I thought it would be interesting to examine
how future higher office ambitions affect the current behavior of members of the
House of Representatives. I was thinking very much in terms of the US. Congress,
and had not thought much about the fact that such an examination would require
a rethinking of some fundamental assumptions. Much to my dismay, I found that
others, notably Hibbing (1986) had already conducted such an examination. But this
led me to think beyond just legislative behavior, and to bring to bear some of the
ideas I got from reading Coase (1937, 1960), North (1981), North and Thomas (1973)
and Williamson (1985), among others. From that perspective, goal-oriented behavior
is both shaped by and shapes institutional rules and structures. If we change our
assumptions about what goals are motivating politicians, then it seems we might
also want to reconsider our predictions about how those politicians will interact with

their institutional environment.

95

So even as I was writing the earlier versions of the essays that make up this
project, I was rethinking what Cox (1997) called “the institutional design question”.
Having recognized that the idea of progressive ambition had its foundation in a more
general way of thinking about political careers, I thought more about how this alter-
native conceptualization of the self-interested politician could yield predictions about
the institutional structure of legislatures. Despite some evidence for the importance
of higher office ambition in understanding the US. Congress, our understanding of
congressional organization is only marginally improved. But an ambition theory of
legislative organization has the potential to be applied generally to legislatures with
relatively fewer adjustments than theories which posit the legislator as single-minded
seeker of reelection.

For legislators, and for politicians more generally, the fundamental goal is prob—
ably best conceived as political survival. Such does not even necessarily entail re-
maining in elective or even appointive office, but rather ‘staying in the game’. Even
if a former Officeholder is relegated to work as a lobbyist, or in a thinktank, or as a
television personality, they still get to experience the give and take of politics, which
for many seems one of life’s necessities. For whatever reasons, whether it is Aris-
totle’s philotimia, or achievement motivation, or power motivation, or an activist’s
desire for good public policy, people get into politics because it fulfils. some need in
themselves.

These are issues that the political scientist finds difficult to address. We can-
not put politicians on the therapist’s couch and have them reveal their innermost
thoughts to us. And even surveys of politicians must be carefully designed in order
to elicit reliable responses, and are rarely administered with any success to anyone
much higher on the career ladder than state legislators. Richard Fenno’s work, uti-
lizing participant observation and personal interviews, probably comes as close as

any in political science to revealing the complicated set of motivations that drive the

96

political personality. But for all its use in informing theory, Fenno’s work (and other
work like it) cannot really be used to test falsifiable hypotheses.

So we must make simplifying assumptions. We assume that individuals are goal-
oriented maximizers of expected utility. Often, we even further assume that goals are
related to the individual’s self—interest. We assume perfect information, we assume
that the actions of different people are chosen sequentially when in reality they are
chosen simultaneously, and vice versa. The list could go on forever. We often know
these assumptions to be false, and yet we make them anyway, because without them,
theory is impossible.

But assumptions should be evaluated. If we can explain much more with a theory
that is only slightly more complicated, we should consider the new theory superior.
But it is not always easy to know where the tradeoff should be made. I have argued
that by changing the fundamental assumption handed down to modern congressional
scholarship by Mayhew (1974), we can substantially improve the generalizability of
the sort of legislative theories pioneered by scholars of the United States Congress.
Congressional scholars are beginning to recognize the need for generalization beyond
one set of rather unique institutions, looking to state legislatures and to parliaments
around the world. But the enterprise should not be to create hundreds of distinct
theories crafted, similarly, to fit one or few sets of institutions, but to return to first
principles and ask ourselves how we can design our theories to apply to more and
more places and times.

Despite our theorizing, legislatures are mixed-goal institutions. We know this
from observation. The question for the political scientist, though, is which of the
myriad goals we might attribute to legislators can safely be assumed away, and which
must be accounted for in theory. My position is that any goals that are related to
careers in public office are worth considering; that, in fact, the added complexity

of moving from the reelection imperative to ambition theory as a foundation is an

97

acceptable tradeoff. Using ambition theory, we can explain everything that theories
built on the reelection imperative can explain and more, and we can do so while
maintaining the convenience of assuming what moral theorists would term an egoistic
perspective. _

This is not to say that I believe no other considerations or goals enter into the
decisionmaking of politicians. Fenno (1973) discussed five goals the pursuit of which
shape legislative behavior. The most important departure from an egoistic framework
in Fenno’s perspective is the concern for good public policy. I agree with Fenno
that politicians often demonstrate such a concern. But it becomes difficult to make
deductively valid predictions when policymaking ceases to be solely instrumental,
from an analytic point of view. Schlesinger’s (1966) typology of ambition allows
for a mixed-goal dynamic without requiring us to discard some of the important
properties of current legislative theories. By accounting for a slightly larger set
of goals and ambitions, I believe we can improve our understanding of the way
legislatures work, and that the value added is worth the increased complexity of our

subsequent theories.

98

Bibliography

Adler, E. Scott. 2002. “Why Congressional Reforms Fail: Reelection and the House
Committee System.” Chicago: University of Chicago Press.

Aghion, Philippe, Alberto Alesina, and Francesco Tfebbi. 2004. “Endogenous Polit-
ical Institutions.” Working Paper, Harvard University Department of Economics.
http://post.economics.harvard.edu/faculty/alesina/papers/aatqje.pdf.

Alchian, Armen A., and Harold Demsetz. 1972. “Production, Information Costs,
and Economic Organization.” The American Economic Review. 62(5): 777-795.

Aldrich, John H. 1995. Why Parties? Chicago: University of Chicago Press.

Aldrich, John H., Calvin C. Jillson, and Rick W. Wilson. 2002. “Why Congress?
What the Failure of the Confederation Congress and the Survival of the Federal
Congress Tell Us About the New Institutionalism.” In Party, Process, and Po-
litical Change in Congress. Eds. David W. Brady and Mathew D. McCubbins.
Stanford: Stanford University Press.

Aldrich, John H., and David W. Rohde. 2000. “The Consequences of Party Organi-
zation in the House: The Role of Majority and Minority Parties in Conditional
Party Government.” In Polarized Politics: Congress and the President in a Par-
tisan Era, eds. Jon Bond and Richard Fleisher. Washington, DC: Congressional
Quarterly Press.

Ambrose, Stephen E. 1987. Nixon: The Education of a Politician, 1913-1962. New
York: Simon and Schuster.

Arrow, Kenneth A. 1951. Social Choice and Individual Values. New York: Wiley.

Bates, Robert H. 2000. “Institutions and Economic Performance.” Prepared for
the Villa Borsig Workshop Series - The Institutional Foundations of a Market
Economy, at the Development Policy Forum of the German Foundation for In-
ternational Development (DSE), Berlin, Germany. February 23-25.

Beck, Nathaniel, Jonathan Katz, and Richard Tucker. 1998. “Taking Time Seri-
ously: Time-Series-Cross-Section Analysis with a Binary Dependent Variable.”
American Journal of Political Science 42: 1260-1288.

Bentley, Eric, ed. 1971. Thirty Years of Treason: Excerpts from Hearings Before
the House Committee on Un-American Activities, 1938-1968. New York: Viking
Press.

99

Bianco, William T. 1984. “Strategic Decisions on Candidacy in US. Congressional
Districts. Legislative Studies Quarterly 9:351-364.

Binder, Sarah A. 1997. Minority Rights, Majority Rule. New York: Cambridge Uni-
versity Press.

Black, Gordon S. 1972. “A Theory of Political Ambition: Career Choices and the
Role of Structural Incentives.” American Political Science Review 66: 144-59.

Bond, Jon R., Cary Covington, and Richard Fleisher. 1985. “Explaining Challenger
Quality in Congressional Elections.” The Journal of Politics 47: 510-529.

Brace, Paul. 1984. “Progressive Ambition in the House: A Probabilistic Approach.”
The Journal of Politics 46: 556-571.

Cain, Bruce, John Ferejohn, and Morris Fiorina. 1987. The Personal Vote: Con-
stituency Service and Electoral Independence. Cambridge: Harvard University
Press.

Carey, John M., and Matthew Shugart. 1995. “Incentives to Cultivate a Personal
Vote: A Rank Ordering of Electoral Formulas.” Electoral Studies 14: 417-39.

Carey, John M., Gary F. Moncrief, Richard G. Niemi, and Lynda W. Powell. 2003.
“Term Limits in the State Legislature: Results from a New Study of the 50
States.” Paper presented at the Annual Meeting of the American Political Sci-
ence Association.

Carson, Jaime L., Michael H. Crespin, Jeffery A. Jenkins, and Ryan J. Vander Wie—
len. 2004. Shirking in the Contemporary Congress: A Reappraisal. Political
Analysis 12:176179.

Clausewitz, Carl Von. 1873. On War Trans. J.J. Graham. London: N. Tfubner.

Cnudde, Charles F., and Donald J. McCrone. 1966. “The Linkage Between Con-
stituency Attitudes and Congressional Voting Behavior: A Causal Model.” Amer—
ican Political Science Review 60: 66-72.

Coase, Ronald H. 1937. “The Nature of the Firm.” Economica, New Series. 4(16):
386-405

 

. 1960. “The Problem of Social Cost.” Journal of Law and Economics.

100

Collie, Melissa P. 1981. “Incumbency, Electoral Safety, and Tumover in the House
of Representatives, 1952-76.” American Political Science Review 75:119—31.

Cook, Thomas D. and Donald T. Campbell. 1979. Quasi-experimentation: Design
and analysis issues for field settings. Boston: Houghton-Miffiin.

Cooper, Joseph, David W. Brady, and Patricia A. Hurley. 1977. “The Electoral Ba-
sis of Party Voting: Patterns and 'Ifends in the US. House of Representatives.”

In The Impact of the Electoral Process, eds. Louis Misel and Joseph Cooper.
Beverly Hills: Sage.

Cover, Albert D. 1977. “One Good Term Deserves Another: The Advantage of In-
cumbency in Congressional Elections.” American Journal of Political Science 21:
523-541.

Cover, Albert D., and David R. Mayhew. 1977. “Congressional Dynamics and the

Decline of Competitive Congressional Elections.” In Congress Reconsidered, eds.
Lawrence C. Dodd and Bruce I. Oppenhemier. New York: Praeger Publishers.

Cox, Gary W. 1987. The Efiicient Secret. Cambridge: Cambridge University Press.

 

1997. Making Votes Count: Strategic Coordination in the World’s
Electoral Systems. Cambridge University Press. New York, NY.

Cox, Gary W., and Jonathan N. Katz. 1996. “Why Did the Incumbency Advantage
in US House Elections Grow?”American Journal of Political Science 40: 478-497.

Cox, Gary W., and Mathew D. McCubbins. 1993. Legislative Leviathan: Party
Government in the House. Berkeley: University of California Press.

 

. 2002. “Agenda Power in the US. House of Representatives, 1877-
1986.” in Party, Process, and Political Change in Congress: New Perspectives
on the History of Congress, ed. David W. Brady and Mathew D. McCubbins.
Stanford, CA: Stanford University Press.

 

. 2005. Setting the Agenda:- Responsible Party Government in the U.S.
House of Representatives. New York: Cambridge University Press.

Crespin, Michael H. 2005. “Congressional Voting Behavior in Response to Changing
District Boundaries.” Paper presented at the Annual Meeting of the Midwest
Political Science Association, Chicago, IL.

Cummings, Milton G. 1966. Congressmen and the Electorate. New York: Free Press.

101

Davidson, Roger H. 1990. “The Legislative Reorganization Act of 1946.” Legislative
Studies Quarterly 15: 357-373.

Denzau, Arthur, William Riker, and Kenneth Shepsle. 1985. “Farquharson and
Fenno: Sophisticated Voting and Homestyle.” American Political Science Re-
view 79: 1117-1134.

Diermeier, Daniel, and Keith Krehbiel. 2003. “Institutionalism as Methodology.”
Journal of Theoretical Politics 15:123-144.

Dies, Martin. 1963. Martin Dies’ Story. New York: Bookmailer.

Dion, Douglas. 1997. Turning the Legislative Thumbscrew. Ann Arbor: University
of Michigan Press.

Dixit, Avinash K. 1996. The Making of Economic Policy: A Transaction-Cost Poli-
tics Perspective. Cambridge, MA.: MIT Press.

Erikson, Robert S. 1971. “The Advantage of Incumbency in Congressional Elec-
tions.” Polity 3: 395-405.

Fenno, Richard F. 1966. Power of the Purse: Appropriations Politics in Congress.
Boston: Little, Brown, and Company.

 

. 1973. Congressmen in Committees. Boston: Little, Brown.

 

. 1978. Homestyle: House Members in their Districts. Boston: Little,
Brown.

 

. 1982. The United States Senate: A Bicameral Perspective. Washing-
ton, DC: AEI Press.

Field, Alexander James. 1981. “The Problem with Neoclassical Institutional Eco-
nomics: A Critique with (Special Reference to the North/ Thomas Model of Pre-
1500 Europe.” Explorations in Economic History. 18: 174-198.

Fiorina, Morris P. 1989. Congress: Keystone of the Washington Establishment, Sec-
ond Edition. New Haven: Yale University Press.

Fowler, Linda L. 1993. Candidates, Congress and the American Democracy. Ann
Arbor: University of Michigan Press.

Friedman, Jeffrey. 1996. The Rational Choice Controversy: Economic Models of

102

Politics Reconsidered. New Haven, CT: Yale University Press.

Froman, Lewis A., Jr., and Randall B. Ripley. 1965. “Conditions for Party Lead-
ership: The Case of the House Democrats.” American Political Science Review
59: 5263.

Froot, K. A. 1989. “Consistent Covariance Matrix Estimation with Cross-sectional

Dependence and Heteroskedasticity in Financial Data.” Journal of Financial and
Quantitative Analysis 24: 333-355.

Gamm, Gerald, and Kenneth A. Shepsle. 1989. “Emergence of Legislative Institu-
tions: Standing Committees in the House and Senate, 1810-1825.” Legislative
Studies Quarterly 14: 39- 66.

Gates, Scott, and Brian D. Humes. 1997. Games, Information, and Politics: Ap-
plying Game Theoretic Models to Political Science. Ann Arbor: University of
Michigan Press.

Gilligan, Thomas, and Keith Krehbiel. 1990. “Organization of Informative Commit—
tees by a Rational Legislature.” American Journal of Political Science 34: 531-64.

Gilligan, Thomas, William Marshall, and Barry Weingast. 1989. “Regulation and
the Theory of Legislative Choice: The Interstate Commerce Act of 1887.” Jour-
nal of Law and Economics 32: 35—61.

Goodman, Walter. 1968. The Committee: The Extraordinary Career of the House
Committee on Un-American Activities. New York: Farrar, Straus, and Giroux.

Gould, Andrew C., and Peter J. Baker. 2002. “Democracy and Taxation.” Annual
Review of Political Science. 5: 87-110.

Green, Donald and Ian Shapiro. 1994. Pathologies of Rational Choice Theory. New
Haven, CT: Yale University Press.

Greene, William H. 1994. “Accounting for Excess Zeros and Sample Selection in
Poisson and Negative Binomial Regression Models.” Stern School of Business,
New York University, Department of Economics. Working Paper No. 94-10.

Griffin, John D. 2003. “Ambition and Legislative Organization.” Unpublished
Manuscript. .

Hasbrouck, Paul D. Party Government in the House of Representatives. New York:
The MacMillan Company.

103

Hayek, Friedrich A. 1973. “Scientism, Historicism and the Problem of Rationality.”
In Modes of Individualism and Collectivism Ed. John O’Neill. London: Heine-
mann .

Herrick, Rebekah, and Michael K. Moore. 1993. “Political Ambition’s Effect on Leg-
islative Behavior: Schlesinger’s Typology Reconsidered and Revisited.” Journal
of Politics 55: .765-776.

Herron, Michael C. 2004. “ Studying Dynamics in Legislator Ideal Points: Scale
Matters.” Political Analysis 12: 182-190. '

Herron, Michael C., and Kenneth W. Shotts “Term Limits and Pork.” Legislative
Studies Quarterly 31: 383-403.

Hibbing, John R. 1986. “Ambition in the House: Behavioral Consequences of Higher
Office Goals Among US. Representatives.” American Journal of Political Sci-
ence 30: 651-65.

Humes, Brian D. 1989. “Congress at the Bicentennial: A Comment.” Legislative
Studies Quarterly 14: 135-45.

Jacobson, Gary C. 1989. “Strategic Politicians and the Dynamics of US. House
Elections, 1946-86.” American Political Science Review 83: 773—93.

 

. 2004. The Politics of Congressional Elections. 6th Ed. New York:
Longman. -

Jacobson, Gary C., and Samuel Kernell. 1983. Strategy and Choice in Congressional
Elections. 2nd ed. Binghamton, N .Y.: The Vail—Ballou Press.

Jenkins, Jeffery A, Michael H. Crespin, and Jamie L. Carson. 2005. “Parties as
Procedural Coalitions in Congress: An Examination of Differing Career 'Ifacks.”
Legislative Studies Quarterly 30: 365-390.

Kazee, Thomas A. 1994. Who Runs for Congress? Ambition, Context, and Candi-
date Emergence. Washington, DC: CQ Press.

Katz, Jonathan N. and Brian R. Sala. 1996. “Careerism, Committee Assignments,
and the Electoral Connection.” American Political Science Review 90: 21-33.

King, David C. 1997. Turf Wars. Chicago, IL: University of Chicago Press.

104

King, Gary, Robert O. Keohane, and Sidney Verba. 1994. Designing Social Inquiry:
Scientific Inference in Qualitative Research. Princeton, NJ: Princeton University
Press.

Krehbiel, Keith. 1987. “Why are Committees Powerful?” American Political Sci-
ence Review 81: 929-35.

 

. 1991. Information and Legislative Organization. Ann Arbor: Univer-
sity of Michigan.

 

. 1993. “Where’s the Party?” British Journal of Political Science 23:
235-66.

 

. 1998. Pivotal Politics: A Theory of US Lawmaking. Chicago: Uni-
versity of Chicago Press. -

Krasno, Jonathan S., and Donald P. Green. 1988.. “Preempting Quality Challengers
in House Elections.” Journal of Politics 50:920—36.

Krasno, Jonathan S. 1994. Challengers, Competition, and Reelection: Comparing
Senate and House Elections. New Haven: Yale University Press.

Lambert, D. 1992. “Zero-inflated Poisson Regression with an Application to Defects
in Manufacturing”. Technometrics 34:1-14.

Lee, Frances E., and Bruce I. Oppenheimer. 1999. Sizing Up the Senate: The Un-
equal Consequences of Equal Representation. Chicago: University of Chicago
Press.

Long, J. Scott, and Jeremy Freese. 2006. Regression Models for Categorical Depen-
dent Variables Using STATA. College Station, TX: Stata Press.

Lublin, David I. 1994. “Quality, Not Quantity: Strategic Politicians in US. Senate
Elections, 1952-1990.” Journal of Politics 56: 228-241.

MacNeil, Neil. 1963. Forge of Democracy. New York: David McKay.

Maltzman, Forrest. 1997. Competing Principals: Committees, Parties, and the Or-
ganization of Congress. Ann Arbor: University of Michigan Press.

Matthews, Donald R. 1959. “The Folkways of the United States Senate: Conformity

to Group Norms and Legislative Effectiveness.” American Political Science Re-
view 53: 1064-1089.

105

 

1960. US. Senators and Their World. Chapel Hill: University of
North Carolina Press.

Mayhew; David R. 1974. Congress: The Electoral Connection. New Haven, Con-
necticut: Yale University Press.

McKelvey, Richard D. 1976. “Intransitivities in Multidimensional Voting Models and
Some Implications for Agenda Control.” Journal of Economic Theory 2: 472-482.

Miller, Warren E., Donald E. Stokes. 1963. “Constituency Influence in Congress.”
American Political Science Review 57:45-56.

Monroe, Nathan W., and Gregory Robinson. 2006. “Do Restrictive Rules Pro-
duce Nonmedian Outcomes? Evidence from Evidence from the 101st-106th Con-
gresses.” Paper presented at the Annual Meeting of the American Political Sci-
ence Association, August 31-September 3, Philadelphia, PA.

Munger, Michael. 1988. “Allocation of Desirable Committee Assignments: Ex-
tended Queues versus Committee Expansion.” American Journal of Political
Science 32:3 17-44.

Nelson, Garrison. 1994. Committees in the US. Congress, 1947-1992, Volume 2.
Washington, DC: CQ Press.

Nixon, Richard M. 1990. RN: The Memoirs of Richard Nixon. New York: Touch-
stone.

North, Douglass C. 1981. Structure and Change in Economic History. New York:
WW Norton.

 

. 1990a. “A Transaction Cost Theory of Politics.” Journal of Theoreti-
cal Politics. 2: 355—67.

-——. 1990b. “Institutions, Transaction Costs, and Exchange.” In Perspec-
tives on Positive Political Economy. Ed. James E. Alt and Kenneth A. Shepsle.
Cambridge, UK: Cambridge University Press.

North, Douglass C., and Robert P. Thomas. 1973. The Rise of the Western World.
Cambridge, UK: Cambridge University Press.

Oppenheimer, Bruce I. 1977. “The Rules Committee: New Arm of the Leadership
in a Decentralized House.” In Congress Reconsidered. Eds. Lawrence C. Dodd

106

and Bruce I. Oppenheimer. New York: Preager.

Perkins, Lynette P. 1981. “Member Recruitment to a Mixed Goal Committee: the
House Judiciary Committee.” T hoe Journal of Politics 43: 348-364.

Persson, Torsten, and Guido Tabellini. 2002. Political Economics: Explaining Eco-
nomic Policy. Cambridge, MA: MIT Press.

Polsby, Nelson W. 1968. “The Institutionalization of the US. House of Representa-
tives.” American Political Science Review 62: 144-68.

 

How Congress Evolves: Social Bases of Institutional Change. New
York: Oxford University Press.

Poole, Keith T., and Howard Rosenthal. 1985. “A Spatial Model For Legislative
Roll Call Analysis.” American Journal of Political Science 29: 357—384.

 

. 1991. “Patterns of Congressional Voting.” American Journal of P0-
litical Science 35: 228—78.

 

. 1997. Congress: A Political-Economic History of Roll Call Voting.
New York: Oxford University Press.

Popper, Karl. 1944. “The Poverty of Historicism, II: A Criticism of Historicist
Methods”. Economica. 11(43):119-137.

—————_—--—. 1945. “The Poverty of Historicism, III”. Economica. 12(46):69—89.

 

1957. The Poverty of Historicism. London: Routledge and Kegan
Paul. ‘

Primo, David, Matthew Jacobsmeier, and Jeffrey Milyo. Forthcoming. “Estimating
the Impact of State Policies and Institutions with Mixed-Level Data.” State Pol-
itics and Policy Quarterly.

Raynor, JD. 1969. “Future Orientation and Motivation of Immediate Activity:
An Elaboration of the Theory of Achievement Motivation.” Psychological Re-
view76:606—610.

 

. 1970. “Relationship Between Achievement-Related Motives, Future
Orientation, and Academic Performance.” Journal of Personality and Social Psy-
chology 15:28-33.

107

Riker, William H. 1955. “The Senate and American Federalism.” The American
Political Science Review 49:452-469.

. 1990. “Political Science and Rational Choice.” In Perspectives on Pos-

itive Political Economy. Ed. James E. Alt and Kenneth A. Shepsle. Cambridge,

UK: Cambridge University Press.

 

Rohde, David W. 1979. “Risk-Bearing and Progressive Ambition: The Case of
Members of the United States House of Representatives.” American Journal of
Political Science 23: 1-26.

 

. 1991. Parties and Leaders in the PostReform House of Representa-
tives. Chicago: University of Chicago Press.

. 1994. “Parties and Committees in the House: Member Motivations,
Issues, and Institutional Arrangements.” Legislative Studies Quarterly 19:341—59.

 

Rohde, David W. 2005. Roll Call Voting Data for the United States House of Repre-
sentatives, 1953—2004. Compiled by the Political Institutions and Public Choice
Program, Michigan State University, East Lansing, MI.

Romer, Thomas, and Howard Rosenthal. 1979. ”Bureaucrats Versus Voters: On
the Political Economy of Resource Allocation by Direct Democracy.” Quarterly
Journal of Economics. 93: 563-87.

Rothenberg, Lawrence S. 2004. “ Much Ado About Very Little: Reply to Herron.”
Political Analysis 12: 191-5.

Rothenberg, Lawrence S., and Mitchell Sanders. 2000. “Severing the Electoral Con-
nection: Shirking in the Contemporary Congress.” American Journal of Political
Science 44: 310—319.

 

. 2004. “Reply to ‘Shirking in the Contemporary Congress: A Reap—
praisal.” Political Analysis 12: 180-1.

Schickler, Eric, and Andrew Rich. 1997. “Controlling the Floor: Parties as Procedu-
ral Coalitions in the House.” American Journal of Political Science 41: 1340-75.

Schlesinger, Joseph, 1966. Ambition and Politics: Political Careers in the United
States. Chicago: Rand McNally.

Selznick, Phillip. 1996. “Institutionalism ‘Old’ and ‘New’.” Administrative Science
Quarterly. 41(2): 270-277.

108

Shepsle, Kenneth A. 1978. The Giant Jigsaw Puzzle: Democratic Committee As-
signments in the Modern House. Chicago: University of Chicago Press.

 

. 1979. “Institutional Arrangements and Equilibrium in Multidimen-
sional Voting Models.” American Journal of Political Science. 23: 27-60.

Shepsle, Kenneth A., and Mark S. Bonchek. 1997. Analyzing Politics: Rationality,
Behavior, and Institutions. New York: WW Norton.

Shepsle, Kenneth A., and Barry R. Weingast. 1981. “Structure-Induced Equilibrium
and Legislative Choice.” Public Choice. 37: 503-19.

 

. 1987. “The Institutional Foundations of Committee Power.” Ameri-
can Political Science Review 81: 85-104.

Simon, Herbert. 1947. Administrative Behavior: A Study of Decision-Making Pro-
cesses in Administrative Organizations New York: The Free Press.

 

. 1957 Models of Man New York: John Wiley.

Squire, Peverill, and Keith E. Hamm. 2005. 101 Chambers: Congress, State Legisla-
tures, and the Future of Legislative Studies. Columbus, OH: Ohio State University
Press.

Stone, Walter J. and L. Sandy Maisel. 2003. “The Not-So-Simple Calculus of
Winning: Potential US. House Candidates’ Nominations and General Election
Prospects.” Journal of Politics 65: 951-77.

Stone, Walter, L. Sandy Maisel, and Cherie Maestas. 2004. “Incumbent Quality as a
Deterrent to Strong Potential House Challengers: Beyond the Strategic Politician
Thesis.” American Journal of Political Science 48: 479-495.

Taylor, Andrew J. “Size, Power, and Electoral Systems: Exogenous Determinants of
Legislative Procedural Choice.” Legislative Studies Quarterly 31: 323-45.

Thomson, J AK 1953. The Ethics of Aristotle. London: Allen & Unwin.

Tsebelis, George. 1990. Nested Games: Rational Choice in Comparative Politics.
Berkeley: University of California Press.

Victor, Jennifer Nicoll. 2005. “ ‘Hey Congressman, Want a Promotion?’: Getting
Busy in the House.” Paper presented at the Annual Meeting of the Midwest
Political Science Association, Chicago, IL. April 7-10, 2005.

109

Walras, Leon. 1954. Elements of Pure Economics. Cambridge, MA: Harvard Uni-
versity Press.

Weingast, Barry, and William Marshall. 1988. “The Industrial Organization of
Congress; or, Why Legislatures, Like Firms, Are Not Organized as Markets.”
Journal of Political Economy 962132-63.

Weissberg, Robert. 1978. “Collective vs. Dyadic Representation in Congress.”
American Political Science Review 72:535-547.

Williams, R. L. 2000. “A Note on Robust Variance Estimation for Cluster-correlated
Data.” Biometrics 56: 645-646.

Williamson, Oliver E. 1985. The Economic Institutions of Capitalism. New York:
The Free Press.

Winter, D. 1973a. The Power Motive. New York: Free Press.

. 1973b. “The Need for Power: 1970.” In Human Motivation. Eds. D.
McClelland and R. Steele. Morristown, NJ: General Learning.

 

 

’. 1980. “Measuring the Motives of Southern African Political Leaders
at a Distance.” Political Psychology 2: 75-85.

 

. 1982. “Motivation and Performance in Presidential Candidates.” In
Motivation and Society.Ed. A. Stewart. San Francisco: Jossey-Bass.

Winter, D. and Stewart, A. 1977. “Power Motivation.” In Dimensions of Personal-
ity. Eds. H. London and J. Ehner. New York: Wiley.

 

. 1978. “Content Analysis as a Method of Studying Political Leaders.”
In A Psychological Examination of Political Leaders Ed. M. Hermann. New
York: Free Press.

Wooldridge, J. M. 2002. Econometric Analysis of Cross Section and Panel Data.
Cambridge, MA: MIT Press.

Zellner, A. 1962. “An Efficient Method of Estimating Seemingly Unrelated Re-
gressions and Tests of Aggregation Bias.” Journal of the American Statistical
Association 57: 500-9.

110