ArticlePDF Available

Partisan Competition and the Decline in Legislative Capacity among Congressional Offices

  • Trinity University

Abstract and Figures

Since the 1990s, members of the U.S. House have shifted official resources away from legislative functions to representational activities. We reveal this decline using an original data set constructed from 236,000 quarterly payroll disbursements by 1,090 member-offices for 120,000 unique House staff between the 103rd and 113th Congresses. The data allow us to examine plausible alternative explanations for this decline, one rooted in the centralization of legislative power over time and the other in conservatives' desires to contract government power. We show that the decline in legislative capacity is symmetrical between and consistent within both parties, contrary to expectations rooted in asymmetrical, ideological sabotage. Additionally, this divestment occurs within incumbent member-offices over time, accelerates when new members of either party replace incumbents, and persists when the out-party majority control changes. We conclude that perpetual competition over institutional control and centralization of legislative functions motivates declining legislative capacity among individual members.
Content may be subject to copyright.
Partisan Competition and the Decline in Legislative Capacity among
Congressional Offices
Jesse Crossona,b, Alexander Furnasc, Timothy LaPirad,e, Casey Burgatf
aPrinceton University
bTrinity University
cUniversity of Michigan
dJames Madison University
eCenter for Effective Lawmaking
fR Street Institute
January 29, 2020
Conditionally accepted, Legislative Studies Quarterly
Abstract. Since the 1990s, members of the U.S. House have shifted official resources away from legislative functions to repre-
sentational activities. We reveal this decline using an original data set constructed from 236,000 quarterly payroll disbursements
by 1,090 member-offices for 120,000 unique House staff between the 103rd and 113th Congresses. e data allow us to examine
plausible alternative explanations for this decline, one rooted in the centralization of legislative power over time and the other
in conservatives’ desires to contract government power. We show that the decline in legislative capacity is symmetrical between
and consistent within both parties, contrary to expectations rooted in asymmetrical, ideological sabotage. Additionally, this di-
vestment occurs within incumbent member-offices over time, accelerates when new members of either party replace incumbents,
and persists when the out-party majority control changes. We conclude that perpetual competition over institutional control and
centralization of legislative functions motivates declining legislative capacity among individual members.
Keywords: Congress; political parties; legislative staff; legislative capacity; Contract with America.
In their classic article, “US Congressman as Enterprise,Salisbury & Shepsle (1981
) provide a rationale
for congressional observers to rethink congressional behavior. e legislature is not a monolithic body with
535 participants, but rather an industry consisting of 535 “loosely coupled” firms, with the members acting
as CEOs. To better understand the congressional economy, the authors argue, legislative scholarship ought
“to incorporate the phenomena of congressional staff systematically with the analysis of Congress rather than
awkwardly appending it to a discussion of congressmen [sic] as discrete individuals” (562-3). is logic was
a straightforward application of institutional theory to the U.S. House: members of Congress have many
competing goals which may best be understood by simply observing how they strategically deploy their
scarce human resources. is observation was consistent with other foundational treatments of Congress,
including how Congress became institutionalized and professionalized by expanding the role of staff (Polsby
1968), how its institutions employ staff to serve members’ political objectives (Fenno 1973,Mayhew 1974),
and how members allocate Washington and home-based staff to match their career stage (Fenno 1978).
Although Salisbury & Shepsle (1981
), Mayhew (1974) and Fenno (1973) suggest that members allocate
their staff to meet their particular goals, they provide little practical guidance regarding
to map members’
goals onto their human resource needs. Nevertheless, a series of recent studies revisit the central importance
of congressional staff to the observed behaviors of members of Congress. ese studies underscore the
ability of staff to influence a variety of important legislative behaviors and outcomes, including voting and
collaboration networks Montgomery & Nyhan (2017), responsiveness to constituents (Hertel-Fernandez
et al. 2019), connections with interest groups and revolving door activity (LaPira & omas 2017,McCrain
2018), and even members’ effectiveness as lawmakers (Crosson et al. forthcoming). Staff serve such a central
role, in fact, that congressional scholars have recently led reform efforts aimed at improving the experience
and expertise of staff in Congress (e.g., LaPira et al. 2020). Yet while this scholarship has underscored
the general value of congressional staff to members of Congress, the temporal limitations of current data
on congressional staff prevent this work from exploring how over-time changes in political context have
themselves altered value of various types of staff to member enterprises. More specifically, current work
is unable to assess how phenomena such as the ascendance of congressional Republicans as a dominant
majority, the rise in competition over majority control (Lee 2016), growing polarization (McCarty et al.
2016,Poole & Rosenthal 2000), and other political trends have systematically altered market conditions in
which members make their investment decisions.
In this article, we investigate how these well-documented changes in congressional politics have in-
fluenced members’ resource allocation calculus. To do so, we introduce a large new dataset of all House
member-office staff, their responsibilities, and their salaries between the 103rd and 113th Congresses. Using
these data, we first document a general decline in members’ investment in legislative operations. We then
investigate how changes in congressional politics drove this decline. We provide evidence that the rise of
insecure majority control of Congress (Lee 2016) and the simultaneous centralization of lawmaking power
within the party apparatus (Curry 2015) has compelled members of Congress to systematically shift their
enterprises’ resources away from legislative endeavors. In fact, we show that members of both parties have
purchased less legislative labor as a percentage of their overall spending during periods of both budget in-
declines, opting instead to invest a larger share in representational labor, such as constituent
service and (especially) public relations.
e paper proceeds as follows. First, we detail the decline in legislative resource allocations in mem-
bers’ personal offices between the 103rd and 113th Congresses, using our original, detailed categorizations
of individual staffer responsibilities over the time period. We note that these investments have declined in
both real terms and as the share of Members Representational Allowances (MRAs) have themselves shrunk.
Second, we lay out two alternative explanations for these declines based on well-documented historical de-
velopments in Congress: symmetrical party competition versus asymmetrical ideological sabotage. ird,
we explain how our extensive new data set of individual staffers’ primary responsibilities is uniquely able to
adjudicate between the two alternatives. Fourth, we use a series of tests to confirm that member divestment
in legislative staff persists under Republican and Democratic majorities, within Republican and Democratic
offices, among newly elected members of Congress, and even among long-standing members of Congress.
We argue that these trends are consistent with shifts in member staffing priorities as a response to centralized
legislative power in an era of insecure majorities—and not conservative sabotage alone. We then discuss the
implications of our findings for legislative studies and congressional reform efforts.
e Decline in Members’ Legislative Operations
Although members may differ in their policy and representational priorities, representatives face similar
demands on their time, represent similarly sized constituent populations, and discharge similar official duties
as members of Congress. All representatives operate under institutionally dictated personal office budget
constraints, set by the Member’s Representational Allowance (MRA) formula, which differ with respect to
travel distance and local cost of living (Brudnick 2019). Nevertheless, given members’ differences in their
goals and priorities, representatives enjoy near full discretion on how to allocate their MRA spending towards
office expenses such as personnel, franked mail, district office rental, and other overhead costs. In fact, there
are effectively just two constraints on their spending. First, House rules forbid members from employing
more than the maximum 18 full time equivalents, plus up to 4 part time equivalents. Second, members are
personally liable for any allowable expenditures exceeding their formula-dictated MRA budget authorizations
to prevent members from overdrawing their allowances.1In practice, then, members are constrained only
by the amount of funding they are allocated.
1For this reason, very few members spend 100% of their MRA allocation. According to recent personal communications with
House Committee on Administration staff on this issue (
Personal communication with House House Committee on Adminis-
tration staff
2019), it is not uncommon for some to publicize this fact as a symbol of their commitment to frugality with taxpayer
dollars. e vast majority of members fail to spend some portion of their MRA, which are retained by the Chief Administrative
Officer as cash reserves; they are not returned to the treasury. Nevertheless, these amounts are typically extremely small, amounting
to just $40,000 on average.
Given the fixed nature of MRAs, a representatives staffing and spending decisions reflect trade-offs faced
by the legislator, since MRA expenditures devoted to one function reduces resources available for others.
us, because of the freedom with which members may spend their funds, representatives’ observed spending
patterns provide insight into how members confront these trade-offs, in pursuit of their individual goals. In-
deed, as Salisbury & Shepsle (1981
) state, “the choice of organizational style may often reflect the member’s
own conception of his or her role and the functional priorities associated with it” (560). More specifically, in
operating their individual legislative “enterprises,” members purchase differentiated labor to meet their goals
of reelection, policy influence, and institutional advancement (Fenno 1973). In their pursuit, members hire
some staff to focus on the actual process of legislating, others to work on casework in the district, and still
others on communications with national and local journalists and through social media. Existing literature
has confirmed the importance of these decisions. For example, members with more experienced policy staff
appear better able to advance their own legislative agenda (Crosson et al. forthcoming), while senior staff
connectedness may enable members to pursue new collaborations (Montgomery & Nyhan 2017). us,
traditionally, members have made use of professional staff as a means for pursue their goals. According to
Salisbury & Shepsles (1981
) logic, then, the relative amounts spent on these types of staff reflect members’
cross-sectional and over-time differences in their priorities as representatives.
In spite of the demonstrated importance of congressional staff and potential for providing insight into
member priorities, political scientists have remained unable to observe members’ investments systematically
over time and in response to changes political context. In this paper, we introduce a unique new dataset of
legislative staff, their responsibilities, and their compensation in the U.S. House, which enables us to observe
these sorts of over-time changes. Using these data, we find that members have systematically divested in
legislative staff — specifically, aides responsible for policy portfolios within their office — over the past two
decades, even as their budgets have been slashed considerably in recent years.
We reveal these trends by analyzing payroll disbursements for 120,000 unique personal staff in the U.S.
House, from the 103rd through 113th Congresses. e full longitudinal data set categorizes 236,000 quar-
terly staffer-member observations. ese data include information on compensation, employer, and job
titles. Our main empirical contribution, however, is to systematically categorize all staffers into five “pri-
mary responsibility” categories—legislative, political management, office management, communications,
and constituency service—which capture the major responsibilities of congressional staff in the U.S. House.
We then collapse these data into three main staffer categories: legislative, representational, and administrative
We systematically identify staff responsibilities using the quarterly
Congressional Yellow Book
that coincide with the quarterly payroll statements. Our protocol permits some staff to have partial roles
in multiple categories, so our process offers a higher degree of precision than simply relying on job titles.
Moreover, relying on job titles alone may hide or inflate domain-specific roles (Petersen 2011), especially as
they relate to legislators’ dual representational and lawmaking responsibilities, which frequently conflict with
each other (Fenno 1973). To the best of our knowledge, our 236,000 detailed coding decisions represent
the largest, most systematic account of member of Congress’s resource allocation decisions ever reported.2
Figure 1 depicts over-time trends in the three broad responsibility categories discussed above. e first
category is
legislative staff
. Staffers were placed into this category based on several factors, summarized in
Appendix B. Most notably, however, staffers were placed in the legislative category if their
Yellow Book
specifically includes a legislative policy portfolio. With this information, we were able to classify not only
“traditional” legislative staff positions (such as Legislative Director and Legislative Assistant), but occasionally
Chiefs of Staffs, Counsel, and Press Secretaries that have clear, substantive legislative responsibilities in the
Vital Statistics on Congress
aggregates staffer counts, but do not associate them with employing offices. e LegiStorm
commercial directory includes payroll disbursements, but is truncated at calendar year 2001 and does not categorize individuals
by their primary responsibilities. A legacy commercial directory published by Congressional Quarterly between 1993 through
2009 does not include payroll disbursements or responsibilities (Montgomery & Nyhan 2017).
103rd 105th 107th 109th 111th 113th
Mean share of MRA
Legislative staff Representational staff Administrative staff
Fig 1: Mean share of member representational allowances allocated to different staff types. Dots show the
mean percentage, while the loess smoother is estimated based on the full distributions. “Representational
staff” is a collapsed category of constituent service and communications staffers, “Administrative staff” is a
collapsed category including political and office management, and “Legislative staff” is any staffer with policy
responsibilities in the Congressional Yellowbook.
office. e second category,
representational staff
consists of constituency service and communications staff.
administrative staff
consists of both “political management” staff—typically the most senior non-
legislative staffers, often a Chief of Staff who lacks a legislative portfolio—and office management, such
as schedulers, information technology personnel, and other support staff.3At this most general level of
analysis, the share of resources spent on administrative staff has sharply declined. More interestingly, we reveal
that members have shifted staffing resources systematically from legislative functions to representational
operations. e share of staffing resources allocated to legislative staffers peaked in the 104th Congress, with
3e original scheme may categorize political managers as also having partial constituency service or communications respon-
sibilities, so portions of their compensation may already be allocated to the
Legislative staff
Representational staff
in Figure 1.
the median office allocating 27.1 percent, or $230,869 for the median MRA. By the 113th Congress this
share had decreased to 18.3 percent, or $225,768 for the median MRA.4 5
While legislative staff expenditures as a share of MRA authorized budgets has decreases consistently over
time, the value in absolute terms fluctuates as the MRA authorizations expand and partially contract in the
112th Congress. Figure 2 plots inflation-adjusted MRA authorizations to all 440 voting and non-voting
member offices. e median spending on legislative staffers peaked in the 111th Congress at $316,245.
Legislative staff expenditures plummet by nearly $90,000 over subsequent Congresses in response to budget
cuts. us, in recent Congresses, members have faced even starker trade-off pressure between legislative
and non-legislative investments. In spite of this fact, as Figure 1 depicts, legislative staff expenditures have
continued to drop. Indeed, in response to MRA budget cuts, members have opted to cut legislative staff
rather than representational staff.6
It is important to reiterate that these data account only for staff in members’ personal offices, compen-
sated with funds allocated by the MRA formula. erefore, it is plausible that the legislative workload has
simply been picked up by the standing committees, ideological caucuses, and legislative support agencies.
However, there is little evidence to support this notion, as the new Republican majority under Speaker Newt
Gingrich followed through on its promises to cut 1/3rd of committee staff, reduce civil service personnel in
the Library of Congress, and eliminate the Office of Technology Assessment (
Vital Statistics on Congress
2018). ese cuts persist through present day. In fact, the only type of committee staff to see an overall
increase in spending has been communication staff, which are primarily oriented toward public relations and
not legislative operations.7Legislative staff levels in committees have remained relatively stable since 2001,
4All salaries are in 2017 inflation adjusted dollars.
5ese results corroborate a more general shift from staff resources shifting from Washington to district offices beginning in
the the 1970s. See Appendix D for the percentage of staff located in Washington and district offices. See Appendix F for models
of several different types of staff beyond legislative staff, including communications and constituency service staff (which comprise
the ”representational staff category).
6is is especially surprising since the “fixed” costs associated with running a congressional enterprise—such as travel and office
rent—ostensibly remain stable.
7See Online Appendix D, Figure A3, page 9.
1995 2000 2005 2010
Total Member Representational Allowance
Fig 2: e rise and fall of Member Representational Allowances. Distributions for 1994 and 1995 in gray
are imputed MRA estimates.
rather than increasing to compensate for individual member divestment.8
Why the Decline: Symmetrical Party Competition or Asymmetric Ideological Sabotage?
Why have members of Congress spent less on legislative staff in recent years? We argue that changes in polit-
ical context—namely, the rise in insecure majorities (Lee 2016) and concomitant centralization of legislating
power in Congress (Curry 2015)—have made investment of resources in legislative staffing less attractive to
8It is unlikely that declining legislative capacity is being augmented by the use of staff whose salaries are shared by committee
offices. For one, few committees’ rules authorize shared staff for all but committee leaders, and for those the House Committee on
Administration limits shared cost amounts between offices. Our data reveal that percentage of salaries paid to shared employees
represents only 1.3% of all payments. It is, therefore, extremely unlikely for such a small subpopulation of shared staffers to make
up for the declines in personal offices legislative staff, because the absolute magnitude of spending on all shared staff is
the magnitude of the decrease in resources devoted legislative staff.
Members of Congress. Given that the rise of insecure majorities coincided with the Republican ascendance
of the 1990s, however, it is also possible that conservatives’ concerted efforts at divesting in the institution
of Congress as a whole redounded to individual members’ investment priorities. us, in the following sec-
tion, we better detail the logic underlying these explanations, drawing upon a series of interviews with both
current and former House staff and Representatives. In doing so, we develop empirical implications that
delineate between the two accounts, which we ultimately examine with our dataset.
How Insecure Majorities Altered Members’ Resource Allocations
As Lee (2016) argues, today’s era of alternating majorities and partisan anxiety—especially in the House—has
increased Congress’s collective focus on reelection. Indeed, given the privileges associated with majority sta-
tus, members realize the value of their co-partisans winning across the country. As a consequence, individual
members delegate more policy development responsibilities to party leaders, who make use of asymmetric
informational advantages (Aldrich 2011,Curry 2015,Koger & Lebo 2017,Lee 2016) and parliamentary
agenda powers (Cox & McCubbins 1993,2005,Sinclair 1998,2011). Party leaders in turn manage party-
differentiated policy agendas and reallocate representational and reelection resources, such as committee
assignments, credit-claiming opportunities, and campaign funds. e result is a legislature with minimal
independent legislative and policy contributions from rank-and-file members. Most major policy changes
are dictated by party leaders (Curry 2015).
We argue that these changes motivate members to spend less of their MRAs on policy aides, such as
legislative directors and legislative assistants who are responsible for tracking, developing, and advancing the
policy priorities of their member-boss. Indeed, with the centralization of the legislative process, members
face receive fewer returns on their investments in high-salaried, long-term, and specialized legislative staffers.9
By the same logic, though, the value of individual members retaining seats in Congress has increased, as
9is is not to say that such staff impart no legislative value to members of Congress, as Crosson et al. (forthcoming) and
others suggest. Rather, the argument rests only on the relative value of such staff declining over time.
majority control in Congress has grown more insecure, thereby increasing the importance of reelection-
oriented activities.
ese dynamics were adeptly summarized in a recent series of interviews with former and current high-
ranking congressional staff. e interviews, conducted between February and June 2017, were adminis-
tered in-person with 52 current senior staff in House and Senate personal, committee, party leadership, and
chamber administrative offices in Washington, DC. Additional interviews were conducted with four former
members of Congress and seven former senior staff from the House and Senate committee offices. Inter-
viewers elicited questions about their personal experiences working in Congress, as well as perspectives on
how offices operate. 10
According to one former Republican member of Congress—a senior appropriator closely aligned with
Newt Gingrich—the decline in member capacity coincided with a concerted effort by party leaders to cen-
tralize legislative control and maximize members’ reelection chances:
[Before the 1994 Republican victory] the schedule was that the average member probably had
his family here, Republican or Democrat.
At Newt Gingrichs insistence when he was Speaker, you leave your families at home because
you’re more likely to get reelected. Politically, it was probably a wise move. Legislatively—for
the good of Congress—it was a disaster, an absolute frigging [sic] disaster, because now their
families, the spouses, the kids are back home. ey go home to politick, presumably do that
[Friday through Monday]. ey’re politicking at home, but their spouses dont want them to
come to Washington so...they come back on Tuesday.
is individual took their assessment a step further, arguing that ”[t]he work product, I think, has shown
that it’s not nearly what it used to be.”
Further still, as a long-standing member who held office both before and after the Contract with America
described, the decline in legislative focus was true for both “Republican or Democrat.” In fact, this position
was a common one among the nearly 60 senior staffers and former members interviewed. For example, this
10Semi-structured interview protocols and additional sampling information may be found in Appendix G. Anonymized tran-
scripts are available upon request.
observation is corroborated by a senior Democratic leadership staffer, who had previously been Chief of Staff
to a member in the mid-2000s:
For years, my old boss had one of the highest retention rates of staff in Congress. en, all of
a sudden, we lost a bunch of staffers [...] We had that tension where we lost two people, one
in DC, one in the district. I could only replace one because I wanted to move some money
around and actually help shore some other people up. It became a tension of do I hire in the
district, where they needed a staffer, or do I hire here?
Ultimately, the office hired only a
staff replacement, despite the fact that this member—a New
England Democrat whose first dimension DW-NOMINATE score falls well left of the party median, and
who still holds a safe blue seat in the House—was comfortably ensconced in the latter stages of his career. A
generation ago, Fenno (1973) suggested a such a member ought to focus on legislative work. Yet the member
and his Chief of Staff concluded that there was little reason to pay a Washington-based legislative staffer.
ese accounts underscore not only how members’ valuation of legislative staff has declined as compe-
tition over majority control heightened, but also how party leaders have encouraged members to focus on
constituent relations and reelection efforts. As previous literature has suggested, party leaders reward those
who obtain valuable resources for reelection (Heberlig et al. 2006,Heberlig & Larson 2007,Kanthak 2007,
Powell 2015). us, in addition to the decreased value of legislative staff, these interviews suggest that parti-
san competition has increased the value of representational staff. Rather than simply connecting the member
with media appearances or performing routine casework, representational staff today face greater pressure to
strengthen the member’s ties with key interests, fundraisers, and constituencies within their districts (Bawn
et al. 2012,Fenno 1978,Miler 2010). As Congress has grown more insecure, the concomitant increased
demand for constituent service and communications forces members to shift resources away from legislative
operations toward representational staff.
In response to these forces, we believe that both parties have improved their ability to build the personal
vote through constituent service (Cain et al. 1987), developed sophisticated public relations efforts Grimmer
(2013), and prioritized messaging bills written largely by party leaders (Curry 2015,Lee 2016). Partisan
warriors in both parties have increasingly and universally de-prioritized their legislative responsibilities as
competition over majority control in both chambers of Congress has increased (Bernhard & Sulkin 2018,
eriault 2013). Given that competition for the majority has remained tight, members of both parties have
refocused attention to representational goals to the detriment of legislative goals.
“Ideological Sabotage” as an Alternative Explanation
As we detail, staffing incentives created by heightened partisan competition offer a compelling explanation for
the coincidental downward and upward trends in share of resources devoted to legislative and representational
staff, respectively, observed in Figure 1. Nevertheless, given that partisan competition is frequently associated
with the rise of the Gingrich speakership and Contract With America, a reasonable alternative explanation for
the observed changes in staffing may derive from simple ideological calculations. at is, perhaps members
allocate staff
to signal their commitment to larger or smaller government, and
to grow (or shrink) legislative capacity in line with their expansionary (or contractionary) ideal preferences.
As a consequence of gaining majority status in 1995, it is therefore possible that Republican members
of Congress unilaterally divested in legislative operations as a matter of ideological commitment. Under
this interpretation, GOP partisans’ divestment in policy-oriented legislative enterprises is driven exclusively
by the “that which governs least” ethos of the conservative coalition. at these institutional changes were
achieved by the
de jure
adoption of the Contract with America when Newt Gingrich became Speaker is the
key piece of evidence to support this claim.
In a recent interview, a senior Democratic leadership staffer—a veteran House aide who has worked in
member, committee, and party leadership offices since the 1990s—characterizes the rhetoric of conservative
members elected in both the 1994 Contract with America class and the 2010 Tea Party members as the
primary cause:
I think [Republicans’] internal and external posture on [reducing legislative operations] was
probably the same. “We need to cut ourselves if we’re going to make sure we’re cutting other
things too.” I really do think it was Tea Party-driven. I think it was the mood of, “We’re spending
too much. If we don’t cut ourselves, we’re no better than the Democrats.” I think it was that
simple. [...]
It was very similar to ’94, was the same idea. “We’re going to cut the Office of Technology
Assessment, we’re going to get rid of all these member service organizations,” and so on. Instead,
“You can do them on your own, but we’re not going to pay for them.
ere were a lot of staffers that went through this, that were in the room when the decisions
were being made. One person told me that they knew it was a bad idea, but they’re doing it
because of the [Tea Party] members...I dont know if you’ll ever get any of them on the record,
or even off the record, to say that. I mean, I think a lot of the long term, dedicated Republican
staff look at the cuts that were made in ’94 and ’95 and then again in 2011 and say, “We’re not
doing ourselves any favors. We know that.
Similar rhetoric is unheard of among Democrats, and certainly cannot be attributed to them, since the
Contract was a Republican initiative. According to this explanation, the sharp decline in legislative resources
inexorably stems from conservative partisan dogma. Despite understanding the ramifications of cuts to
legislative operations staff—or perhaps because of it—members of the Republican Party viewed the costs in
legislate matters to be far below the electoral advantage gained from cultivating an image of fiscal frugality.
Hypotheses and Tests
While the popular narratives we document in our interviews support both the “partisan competition” and
“ideological sabotage” explanations for the decline in legislative staff investments, we can use our detailed
data on members’ allocation of scarce human resources to more systematically adjudicate between the two
explanations. On balance, we believe that partisan competition—and the changes in the relative values
of staffers in different roles it induces—best explain why members of Congress have appeared to divest in
legislative staff. Here, we articulate three sets of tests to adjudicate between these two theories for staff
allocation and change, ultimately finding the most support for the partisan competition explanation.
Perhaps most importantly to the ideological sabotage theory, we should observe asymmetric divestment
from legislative staffing between Republican and Democratic personal offices, as a consequence of ideological
differences between the parties. All else equal, divestment should occur in the post-Contract era and chiefly
among Republicans. Conversely, partisan competition suggests symmetric divestment, as members of both
parties respond equally to the concentration of legislative power in leadership offices and the increasing
representational demands of electoral competition.
is fundamental difference leads to several testable hypotheses about the timing and partisan differences
in legislative staffing trends. Our first set of such hypotheses concern the
in legislative staff resources
over time, and potential partisan differences in this decline.
H1: All legislators entering Congress in the post-Contract with America period devote a smaller share
of their MRAs to legislative staff than legislators that first entered Congress before the Contract.
H1A: Among legislators entering Congress in the post-Contract with America period, Republicans de-
vote a smaller share of their MRAs to legislative staff than Democrats.
We test H1 by interacting legislator party with 1) a dummy variable for whether the legislator was
elected pre/post Contract with America, and 2) a linear time trend for the year the legislator was
We focus on when legislators first entered Congress to capture the era in which they were socialized into
norms of legislating. Freshman members are offered training by party leadership, House Administration
Committee staff, and the Congressional Research Service, including manuals with model budgets and staffing
allocations. Consequently, we expect staff allocations to remain sticky and constrained by members’ pre-
existing conceptions about their legislative roles and responsibilities. As such, we expect divestment —
either due to increased partisan competition or ideological sabotage — to be particularly pronounced among
legislators that are first elected under the new institutional conditions arising after the Contract with America.
In both cases, a significant
with legislator party would support H1 and the “ideological sabotage”
theory—indicating that divestment patterns differ significantly between the two parties.
Next, we investigate whether the incentive to divest from legislative staff is driven by whether or not a
member belongs to the majority party. Given power asymmetries between majority and minority members
in the House, investing heavily in legislative endeavors may make little sense for members of the minority
party. However, if the ideological sabotage hypothesis is correct, the 2006 Democratic wave election should
reverse the divestment trend among Democrats. at is, now in a position to legislate a broad, progressive
policy agenda—including what would become the Affordable Care Act—the asymmetric hypothesis implies
post-2006 Democrats should re-invest in legislative staff. Conversely, the influence of partisan competition
should not be sensitive to which party actually holds the majority at any given point in time.
H2: Minority party members should divest more extensively from legislative staff than do majority party
members under both Republican and Democratic control.
If the ideological sabotage theory is correct, but clear partisan differences are obscured by majority control
of the chamber, we should instead expect the following:
H2A: When Democrats are the party of the majority, Democratic members should re-invest more in
legislative staff than minority Republicans.
We test H2 and H2A by subsetting the panel according to the party of the majority and again estimating
models to interact legislator party and year of first election. If divestment were in fact ideologically motivated
— meaning that Democrats were only divesting while Republicans were in control because of low expected
utility of legislative work when in the minority — we should expect to see Democrats reinvest in legislative
staff once they are in control. In this case, the absence of a partisan difference in divestment trends when
Democrats are in power would lend support to the partisan competition theory.
Our first several tests focus on legislative divestment based on when they are first elected to Congress.
is operationalization captures how members are socialized into the Congress under particular institutional
circumstances. However, we also expect that long-serving members will be somewhat responsive to the same
contextual factors faced by newly elected members. We therefore use a within-member panel design to
investigate how legislative investment patterns fluctuate over time and between the two parties. Here again,
persistent legislative divestment over time would be consistent with the partisan competition theory, while
partisan differences in divestment would be consistent with ideological sabotage:
H3: Members of Congress invest a smaller share of their MRAs in legislative staffing in later Congresses,
regardless of party.
We test this hypothesis by leveraging within-member variation in staff investments, modeling the share of
MRAs directed towards legislative staff with member-level fixed effects. e primary independent variable of
interest is a linear time trend for the 103rd-113th Congresses in our panel. A significant negative coefficient
on this parameter indicates that individual members have shifted their MRA away from legislative staff
during our period of observation, supporting the partisan competition theory. In other words, legislative
divestment over time is not simply a “replacement effect” brought about by large influxes of new members
to Congress. Rather, even long-serving members would adapt over time the changing political context, just
as in the New England Democrat example above.
By contrast, ideological sabotage implies that we should observe within-member declines primarily
among Republican members of Congress. us:
H3A: Republican members of Congress invest a smaller share of their MRAs in legislative staffing in
later Congresses than Democratic members do.
In order to asses H3A, we interact the Congress time from H3 with legislator party.
Finally, we examine the extent to which newly elected members affect overall institutional divestment.
Here, we subset our panel to include only the first terms in office for each member first elected between
1992 and 2012. We use these cross-sectional data to assess whether
members elected before the
Contract with America allocated more of their MRAs to legislative staff than did those elected afterward.
Given the changes associated with the rise in partisan competition, we expect:
H4: Freshman members in later Congresses devote a smaller share of their MRAs to legislative staff than
freshman members in earlier Congresses.
Conversely, according to the ideological sabotage logic, we should observe this effect primarily among
Republican members of Congress:
H4A: Among Freshman members Congress in the post-Contract with America period, Republicans
devote a smaller share of their MRAs to legislative staff than Democrats.
We test H4 using both a linear trend for the year in which the freshman member was first elected, and
using a binary variable for whether the freshman member was first elected before or after the Contract with
America. We then test H4Aby interacting both of these measures with a dummy variable for legislator party.
As briefly introduced above, our Congressional staff allocation panel covers the 103rd-113th Congresses. e
data exclude the 109th Congress for reasons beyond our control.11 Data for the 103rd-106th Congresses
were drawn directly from archived
Statements of Disbursement of the U.S. House
, while data from the
107th-113th Congresses is provided by disbursement records digitized by the DC-based firm LegiStorm.
To categorize individual staffers’ responsibilities, a large team of research assistants used individual staffer
entries in quarterly
Congressional Yellow Book
directories to investigate staffers’ primary work functions.
Congressional Yellow Book
is a longstanding and trusted commercial directory used by Washington
elites, such as lobbyists and congressional staff, since at least the 1970s. ese volumes provide a wealth
of information that are useful for rendering accurate labor responsibility classifications, particularly as they
pertain to legislative responsibility. First, for each member, the books detail which staff serve as “key aides”
in both Washington and district offices. Not only is the staffer’s physical location highly useful (though not
dispositive) for inferring staffer responsibilities, but staffers’ presence in the volume itself provides context
11Due to well-established data quality issues in House disbursement records, there are no informative staff titles included in
2005 or 2006 House records. is made categorizing staffers in the 109th Congress impossible.
for the their responsibilities. In addition, the volumes detail the legislative issue portfolios for relevant staff.
e accuracy of this information is crucial for the creators of the volume, because lobbyists and other ma-
jor consumers use the books to strategically target staff within congressional offices. Finally, the volumes
occasionally list more informative job titles compared to what is reported in disbursement records. is
information is especially useful for classifying staff whose payroll titles are uninformative in payroll records
It is important to underscore here that these data seek to capture a staffer’s
responsibilities, and
not her
responsibilities, as staff undoubtedly take on multiple roles within their offices. However,
we seek to classify staff according to the office function that best describes their role. In cases where no such
function appears to predominate, we take steps to split a staffer’s salary between competing responsibility
categories. However, we do so only when original archive sources provide clear evidence of multiple office
functions, or of job title changes or promotions within the office.13
e detailed responsibility information in these directories is incorporated systematically into our coding
protocol, found in Appendix B.14 e protocol ensures that our data provide not only precise information
about office responsibilities, but also consistent information across offices and throughout the period of
observation. Combining this material with staffers’ job titles is markedly more accurate than relying solely
on job titles for classification, as such titles are not consistently operationalized across offices, personnel,
and historical time period. As we highlight in greater length in Appendix B pages 3 to 5, projects relying
on fully automated coding procedures are likely to yield inaccurate and biased staff allocation measures for
congressional offices. First, as Petersen (2011) notes in the largest survey of staff titles and responsibilities
conducted to date, job titles are quite variable across member offices. Staff assistants in one office may serve
12For example, some members like Rep. Dave Camp list all staff as “Staff Assistants.”
13We argue that interns do not present a problem for our ultimate analysis. First, interns are rarely assigned to policy respon-
sibilities. Second, on the rare occasions they might be assigned to legislative duties, they should be captured in the Yellow Book.
Finally, during our period of observation most interns earn no compensation; the total intern compensation is just 0.4% of all
staff compensation.
14Supplemental Information, pages 6 and 7.
as low-level office administrators, for example, while similarly titled staffers in other offices serve as the lead
policy analysts.15 In fact, even
an office, similar titles do not always indicate similar responsibilities.
Our coding process captures this kind of nuance, particularly as it pertains to legislative staff, as the
record legislative responsibilities for staffers
regardless of staffer title or time period.
is kind of consistency is especially important when attempting to explain allocation patterns over time.
We therefore believe that our coding scheme is both more externally valid since job titles may be inflated for
idiosyncratic reasons and internally reliable over time. Were such errors randomly distributed across offices
and Congresses, they may simply introduce measurement noise, rather than systematic bias. However, this
is unlikely to be the case. In informal interviews conducted for this data collection effort, staff have insisted
that titling conventions have changed appreciably over our period of observation. For example, the military-
inspired title “Chief of Staff” did not grow in popularity until the 1970s, and did not reach its present
usage level until well into the 2000s. Instead, members titled top staffers as “Administrative Assistants.”
Under modern-day naming patterns in the House, such a title is far more commonly associated with low-pay
administrative work than the political management responsibilities of a Chief of Staff. Idiosyncrasies like this
example introduce at least two different kinds of systematic bias. First, when high-seniority members choose
not to adopt new naming conventions, these differences in title conventions will introduce systematic cross-
sectional bias between senior and junior members of Congress. Second, members who
adopt evolving job
titling conventions introduce systematic temporal bias between earlier and more recent congressional offices.
ese biases are not the only potential biases introduced by automated procedures; but, given that we purport
to uncover an over-time decline in member-level legislative investment, they underscore the importance of
our careful, granular operationalization of staffer responsibilities. Here again, the
Yellow Books
are invaluable
15is ambiguity is particularly crucial in coding Chief of Staff. In some offices, the Chief of Staff serves as lead legislative
counsel, effectively assuming the role of Legislative Director. In others, the Chief of Staff focuses primary on maintaining good
relationships in the district or serving as the primary gatekeeper for access to the member. Ensuring accurate coding of chiefs of
staff is of first-order importance, as the salaries of these individuals commonly occupy a large portion of a member’s total spending
on staff.
to the accuracy of our coding procedure, as they record staffers’ legislative responsibilities over the period
of observation. We maintain a high level of confidence in the accuracy of the information provided by
the volumes, as this information is marketed commercially to well-connected policy advocates and other
government relations personnel who rely upon its veracity.
Using these resources, staffers were initially sorted into one of five functional categories:
legislative staff
political management staff
office management staff
communications staff
, and
constituency service staff
Legislative Staff
are staff whose primary responsibilities are to advise the member of Congress on matters
pertaining to policymaking and the legislative process. Responsibilities may include drafting new bills, de-
ciphering legislative language, offering voting or co-sponsorship advice, providing parliamentary procedure
and legislative negotiation expertise, or interacting with stakeholders on behalf of the member.
Management Staff
are staff whose primary responsibilities are to manage the member’s relationships with
other elites in Washington, such as leaders of political parties and issue caucuses, lobbyists, and major donors.
Political Management Staff
deal in clerical responsibilities, such as coordinating office space, materials, and
information technology, bookkeeping, arranging member travel, and reserving meeting space for constituent
Communications Staff
interact with the media on the member’s behalf, including managing social me-
dia, writing press releases, scheduling television appearances, drafting speeches, and submitting op-eds to
newspapers. Finally,
Constituency Service Staff
deal primarily in relations with the member’s constituents.
Most often located in district offices, their responsibilities include handling bureaucratic casework, answering
constituents’ phone calls and mail, and alerting the member to public events in the district.
While some conceptual overlap between these categories undoubtedly exists, they are designed to capture
Mayhews (1974) primary member activities for reelection (advertising, credit-claiming, and position-taking)
and Fennos (1973) member goals (reelection, influence in the chamber, and good public policy). Note
that on occasion a staffer may have more than one title associated with their names. In those cases, after
consultation with the
Congressional Yellowbooks
, these staffers’ responsibilities are split evenly between two
(or more) categories. Consequently, these data represent the most complete, detailed, and accurate account
of individual staffer functions ever compiled.
We connect individually categorized staff with their salary disbursements to create aggregate, office-
level measures of staff allocation over time. For each Representatives’ office, we calculate the share of their
MRA spent on each of the five staffing categories in each year.16,17 e result of this aggregation is an
unbalanced panel dataset of 1,090 members over nine Congresses, for a total of 4,256 observations, based
on 236,000 quarterly individual staffer responsibility coding decisions. We treat salary allocations across
work responsibilities as members’ revealed preferences for organizational priorities. In the analyses to come,
we use the percentage of a representative’s MRA spent on salaries for staffers coded as having primarily
legislative responsibilities as our main dependent variable.
Other Variables
To account for well-known institutional features and personal characteristics that may impact members
behavior, we include a variety of independent variables from the Volden & Wiseman (2014) Legislative
Effectiveness Project. First, research on the “political life cycles” of members of Congress suggests that mem-
bers’ may emphasize the pursuit of policy goals later in their career as they feel safe in their seats and seek
a legislative legacy (Fenno 1978,Hibbing 1991). To account for this we include
—number of
Congresses served at a given point in time—as an independent variable in our models. We also include the
square of seniority to capture the decreasing marginal effect of seniority. Second, we include binary variables
for serving as a
Committee Chair
, a
Subcommittee Chair
, or as a member of
Power Committee
, as these
16Because of missing data, spending for some years are estimated with projections based on quarterly disbursements for one,
two, or three quarters.
17Additional information about the collection, cleaning and backwards projection of MRA totals is available in Appendix A,
pages 1 and 2.
members may be especially well-situated to substitute their personal office legislative staff with committee
resources (Fenno 1973,Patterson 1970,Salisbury & Shepsle 1981
ird, we include the folded 1st-dimension of DW-NOMINATE to account for member
as Volden & Wiseman (2014) suggest that members farther from the chamber median are less legislatively
effective, implying they may invest less in legislative staff. Fourth, because Volden et al. (2013) suggest that
female legislators engage in legislative activity differently than their male counterparts, we include a binary
variable for a member’s gender. is term allows for the possibility that these differences extend to members’
investment patterns.18 Finally, because members in more competitive districts may face differential electoral
incentives that induce them to invest more in communications or constituent service over legislative staff,
we include variables for
Vote Share
and its square. 19
We develop four relevant tests to discriminate between the ideological sabotage and partisan competition
hypotheses. First, we examine whether pre- and post-Contract members from both parties, or only the Re-
publican party, exhibit a reduction in legislative staff investment following the 1994 Republican takeover.
Second, we investigate whether members elected more recently invest less in legislative staff, and whether this
trend is unique to Republicans. ird, we examine whether Democrats reinvest in legislative capacity once
they regain control of the House—or whether they continue the trend of divestment. After demonstrating
that the partisan competition hypothesis appears to best explain the observed trends in the data, we show
that freshmen elected more recently invest far less in legislative staff than freshmen in earlier Congresses,
suggesting that Congress’s perpetual campaign has attracted members who are more inclined to cede legisla-
18Due to the potential for collinearity between race and party in the House, we have excluded race variables in these models.
However the results presented here are robust to the inclusion of indicator variables for Latinx and African-American ethnicities.
19Electorally safe and ideologically extreme members may represent the same basic group of individuals, though standard tests
for collinearity indicate that the variables are not unduly inflating variance. In fact, the coefficients point in the same directions
in the Volden et al. (2013) analysis of gender and legislative effectiveness.
tive endeavors to other actors in Congress. is is consistent with the theory that contemporary congres-
sional context of insecure majorities and the centralization of legislating within leadership requires legislative
staffers fewer high-value skills, which has induced members to divest from legislative staff. Finally, in a sup-
plementary analysis in Appendix F, we investigate whether representational staffers (communications and
constituent service staff) exhibit similar divestment patterns to legislative staff. We find that instead, mem-
bers of both parties have increased their investments in representational staffers over time, which is again
consistent with the symmetric partisan divestment story.20
Legislative Investment and the Contract with America
In our first series of tests, we examine whether members elected after the Republican Contract with America
takeover systematically invest less in legislative staff than members initially elected and professionally social-
ized to legislative productivity norms in earlier Congresses. According to the ideological sabotage hypothesis,
members elected under Gingrich’s Contract with America agenda were committed to cuts in legislative ca-
pacity. However, because these ideological commitments were held only by Republicans, election during the
Newt Gingrich era should decrease legislative investment only among Republicans. Conversely, the partisan
competition hypothesis suggests that both parties should exhibit declines legislative investment over time.
We test these competing predictions directly in Models 1 and 2 of Table 1. In these models, we regress a
member’s share of MRA allocated to legislative staff on a binary variable, indicating whether a member en-
tered Congress before or after the Contract with America. To examine whether or not election post-Contract
was particularly consequential for Republicans, we interact a partisan indicator (with Democrat = 1) with
the post-Contract variable. If the asymmetric hypothesis is correct, we should observe a significant, posi-
tive coefficient on this variable (large enough to overcome negative main effects)—indicating that election
post-Contract matters differentially for Republicans and Democrats. In both Models 1 and 2, as well as
20We thank an anonymous reviewer for suggesting this additional test.
all remaining models presented hereafter, we employ beta regression estimated via MLE, as our outcome
variable is a proportion that cannot take a negative value or a value greater than 1.
Table 1: Legislative Investment Before and After Contract with America
Percentage of MRA Spent on Legislative Staff
Model 1 Model 2 Model 3 Model 4
Elected Post-Contract
0.197∗∗ 0.221∗∗
(0.048) (0.042)
Year of Election
0.019∗∗ 0.023∗∗
(0.005) (0.004)
0.048 0.0680.028 13.848∗∗
(0.033) (0.030) (0.037) (5.669)
Post Contract
Year of Election
0.0250.0250.012 0.009
(0.013) (0.012) (0.018) (0.018)
20.0020.0020.003∗∗ 0.002∗∗
(0.001) (0.001) (0.001) (0.001)
Committee Chair
0.192∗∗ 0.196∗∗ 0.218∗∗ 0.229∗∗
(0.064) (0.063) (0.068) (0.066)
Subcommittee Chair
0.044+0.046 0.038 0.035
(0.026) (0.028) (0.037) (0.031)
Power Committee
0.037 0.036 0.047 0.043
(0.032) (0.030) (0.034) (0.031)
0.273∗∗ 0.255∗∗ 0.1330.057
(0.068) (0.072) (0.055) (0.062)
0.0470.0480.026 0.036
(0.021) (0.019) (0.016) (0.018)
Vote Share
0.028∗∗ 0.028∗∗ 0.029∗∗ 0.028∗∗
(0.006) (0.006) (0.006) (0.006)
Vote Share
20.0002∗∗ 0.0002∗∗ 0.0002∗∗ 0.0002∗∗
(0.00004) (0.00004) (0.00004) (0.00004)
Constant 1.930∗∗ 1.914∗∗ 36.504∗∗ 44.753∗∗
(0.300) (0.292) (10.103) (9.503)
Observations 4,256 4,256 4,256 4,256
R20.066 0.066 0.094 0.098
Log Likelihood 4,312.121 4,313.111 4,389.562 4,399.188
∗∗p<.01; p<.05; +p<.1
As Table 1 illustrates, election during the post-Contract era is negatively associated with member-level
investment in legislative staff. Holding all other variables at their means or optimal values, estimates from
Model 1 predict that the typical pre-Contract members spent 26 percent of their MRA on legislative staff,
compared to the predicted 22.4 percent for post-Contract members. e difference of 3.6 percent amounts
0 1
Elected Before/After Contract with America
% MRA Dedicated to Legislative Staff
Rep Dem
(a) Binary Independent Variable: Model 2
1940 1950 1960 1970 1980 1990 2000 2010 2020
Year First Elected
% MRA Dedicated to Legislative Staff
Rep Dem
(b) Continuous Independent Variable: Model 4
Fig 3: Interactions between Member Party and Recent Election
to a reduction in spending of $41,66921—roughly equivalent to the salary of a one additional full-time
legislative assistant.
ese differences notwithstanding, Republicans do not appear to be noteworthy in their divestment in
legislative staff. Figure 3 depicts the interaction between post-Contract status and partisanship. As the figure
clearly illustrates, there are no significant differences between Republicans and Democrats with respect to
the relationship between post-Contract status and legislative investment.
None of these findings are fully consistent with the asymmetric sabotage narrative, though relying on an
interaction between two binary variables is a difficult test to interpret. us, we introduce an alternative test
in Models 3 and 4 that replace the post-Contract indicator with a nominal variable representing a member’s
first year of election to Congress. Unlike the binary indicator, this variable allows for the possibility that
members elected just before the 1994 wave may not have been substantially different from those elected
immediately after it. Moreover, it allows us to interrogate whether the over-time relationship between the
21Using the average MRA ($1,157,460) in the dataset
election of new members and legislative investment differs between Republicans and Democrats.
As with Models 1 and 2, Models 3 and 4 exhibit a strong, negative relationship between recent election
and legislative investment: each two-year election cycle is associated with a 0.7 percent decrease ($8,102,
predicted at average MRA) in legislative staff spending compared to the previous year.22 Unlike in Models 1
and 2, however, Models 3 and 4 initially appear to provide some modest support for the ideological sabotage
hypothesis: while
Year of Election
is negatively and significantly associated with legislative investment, this
association is less strongly negative among Democrats. Nevertheless, this difference is substantively small.
In fact, as predicted values in Figure 3 indicate,
Year of Election
remains negative for both and differences
between Republicans and Democrats are substantively small. In these models, Democrats may be slightly
less committed to divestment than Republicans, but hardly enough to substantiate claims of asymmetric
In Appendix F in the supplemental material we re-estimate models 1-4 on the share of their MRAs
that Members of Congress allocate to to communications staff and constituent service staff, respectively.
Where we see bipartisan secular declines in legislative staff during the post-Contract era, we see
these two staffing categories. e more recently members are elected, the higher a share of their MRAs they
allocate to staffers with direct representational duties. ese results, detailed in the supplemental materials,
are consistent with the party competition theory.
Legislative Investment and Majority Control
Next, we leverage Democrats’ return to power in 2006 as a means for discriminating between the asymmetric
sabotage and party competition hypotheses. One possible objection to the above analysis is that Democratic
members between 1995 and 2006 may simply have divested due to their minority status. Given power
asymmetries between majority and minority members in the House, investing heavily in legislative endeavors
22Note that, while the coefficient on
appears inordinately large in Model 4 (and in Models 8 and 10, below) this is
only due to the fact that
is interacted with a year (e.g. 2002), which is much larger than the 0,1 dummy variable.
may make little sense for members of the minority party. However, if the ideological sabotage hypothesis is
correct, the 2006 Democratic wave election should reverse the divestment trend among Democrats. Now
in a position to legislate a broad policy agenda, the asymmetric hypothesis implies post-2006 Democrats
should re-invest in legislative staff.
In Table 2, Models 5 and 6 fit similar specifications on data drawn only from Congresses for which
Republicans controlled the House majority, while Models 7 and 8 focus on those with Democratic majority.
Sub-setting the data in this fashion allows us to assess whether members differentially invest in legislative staff
based on their party, holding majority control constant. Alternatively, if Democrats did in fact reinvest in
legislative staff after regaining the majority, then Republican members should now remain the only members
who continue to divest in legislative staff.
When Republicans control Congress,
Year of Election
shows similar relationships as it does in the full
regressions. As shown in Figure 4, when Democrats control Congress,
Year of Election
still negatively and
significantly predicts legislative investment. More importantly, the association between
Year of Election
and legislative investment is not significantly different between Democrats and Republicans. Again, the
evidence points toward the partisan competition hypothesis, as each new cohort from both parties continue
the divestment trends initiated by their post-Contract colleagues.
We introduce member-level fixed effects in Models 9 and 10 to illustrate this point, leveraging only
within-member variance. In place of
Year of Election
, we include a simple
time trend. Doing so
tests whether majority-party Democrats invested in legislative endeavors more than majority-party Repub-
licans and allows the model to hold constant district-level factors that may influence a member’s decision to
invest in legislative staff. Previous research has suggested that different types of constituencies generate de-
mand for different types of representation (Butler 2014,Foster-Molina 2018,Harden 2013), and inclusion
of member-level fixed effects allows the model to hold such factors relatively constant (with the exception of
Table 2: Majority Control and Legislative Investment
Percentage of MRA Spent on Legislative Staff
Model 5 Model 6 Model 7 Model 8 Model 9 Model 10
Year of Election
0.024∗∗∗ 0.026∗∗∗ 0.007∗∗∗ 0.010∗∗∗
(0.006) (0.007) (0.001) (0.002)
0.036 9.412 0.091∗∗∗ 11.232∗∗∗ 0.003 0.140
(0.049) (5.887) (0.028) (3.528) (0.004) (0.091)
Year of Election
0.005 0.006∗∗∗
(0.003) (0.002)
Congress Trend
0.007∗∗∗ 0.008∗∗∗
(0.0004) (0.001)
0.002 0.002 0.007∗∗∗ 0.007∗∗∗
(0.009) (0.013) (0.001) (0.001)
20.002∗∗∗ 0.002∗∗∗ 0.0003∗∗∗ 0.0003∗∗∗
(0.0004) (0.0004) (0.0001) (0.0001)
Committee Chair
0.169∗∗∗ 0.205∗∗∗ 0.160∗∗∗ 0.116∗∗∗ 0.027∗∗∗ 0.026∗∗∗
(0.029) (0.046) (0.010) (0.002) (0.006) (0.006)
Subcommittee Chair
0.010 0.024 0.026 0.008 0.004 0.004
(0.022) (0.024) (0.020) (0.025) (0.003) (0.003)
Power Committee
0.007 0.007 0.0250.023 0.003 0.003
(0.009) (0.018) (0.014) (0.016) (0.004) (0.004)
0.152∗∗ 0.1050.026 0.057∗∗∗ 0.024∗∗ 0.017
(0.067) (0.059) (0.019) (0.017) (0.010) (0.011)
0.048∗∗ 0.055∗∗∗ 0.024 0.022 0.001 0.001
(0.019) (0.020) (0.018) (0.020) (0.005) (0.005)
Vote Share
0.018∗∗∗ 0.017∗∗∗ 0.034∗∗∗ 0.033∗∗∗ 0.004∗∗∗ 0.004∗∗∗
(0.004) (0.003) (0.002) (0.001) (0.001) (0.001)
Vote Share
20.0001∗∗∗ 0.0001∗∗∗ 0.0002∗∗∗ 0.0002∗∗∗ 0.00003∗∗∗ 0.00003∗∗∗
(0.00003) (0.00002) (0.00001) (0.00001) (0.00001) (0.00001)
Constant 46.089∗∗∗ 51.169∗∗∗ 10.810∗∗∗ 17.017∗∗∗ 0.863∗∗∗ 0.944∗∗∗
(11.039) (13.362) (2.007) (4.259) (0.057) (0.078)
Member-Level FE? N N N N Y Y
Majority Restriction R R D D All All
Observations 2,974 2,974 861 861 4,256 4,256
R20.128 0.129 0.049 0.052 0.554 0.554
Log Likelihood 3,046.756 3,049.511 1,134.396 1,135.698
+p<0.1; p<0.05; ∗∗p<0.01
Note: Models 7 and 8 drop
2, due to collinearity problems with
Year of Election
these models also do not feature clustered standard errors, due to estimation issues
1960 1970 1980 1990 2000 2010
Year First Elected
% MRA Dedicated to Legislative Staff
Rep Dem
(a) Republican Controlled Congresses
1960 1970 1980 1990 2000 2010
Year First Elected
% MRA Dedicated to Legislative Staff
Rep Dem
(b) Democratic Controlled Congresses
Fig 4: Legislative Investment by Congress for Republicans and Democrats
redistricting years).
Figure 5 shows that the time trend is strongly negative, and is not statistically distinguishable between
Republicans and Democrats, even when such effects are included. at is, even if we hold the individual
member—and, to some extent, district—constant, investment in legislative staff decreased over time, regard-
less of party. is finding is consistent with the symmetrical partisan competition view. Additionally, this
finding is consistent with H3, which states that even long-serving members should adapt their investment
patterns to the changing political context. Indeed, as the negative coefficient on the time trend indicates,
members are significantly divesting in legislative staff over time—even when we leverage only within member
ese analyses reveal that divestment in legislative operations are not unique to Republicans. e impli-
cation is that Speaker Gingrich was most likely not as instrumental for broader divestment trends in Congress
as some assume. Instead, we infer that Gingrich coincidentally came to power at a time when broader elec-
toral conditions empowered party leaders, making individual legislative enterprises less important.
103 105 107 109 111 113
% MRA Dedicated to Legislative Staff
Rep Dem
Fig 5: Member-level fixed effects
Even still, it remains possible that the Republican takeover in 1995 accelerated the divestment process.
Moreover, we recognize these results may imply that Gingrich indeed aggressively encouraged divestment
in legislative enterprises, as suggested by his long-time ally that we interviewed. In Figure 6, we report the
results of a series of regressions that use the basic specification found in Model 1, as well as Congress-level
fixed effects, to more accurately observe replacement effects as new members are elected to the House over
time. Across the models, we replace the binary pre-/post-Contract variable with a pre-/post-binary variable
for each year of the first election listed in the x-axis. e coefficient and its corresponding 95% confidence
interval are reported in the figure, with a reference line for 1994.
As the figure demonstrates, not all cutoff variables take on a negative value. In fact, it is not until coeffi-
cients reported for members elected in 1992 that the cutoff variable takes on a negative, albeit insignificant,
sign. e following two years, 199323 and 1994, however, are not only negatively associated with legislative
investment but approach statistical significance (p0.0503 and 0.0501). at is, legislators elected during
23ere were five special elections in 1993.
Fig 6: Estimated Coefficients for Electoral Cutoff Variables
Contract With America
1971 1975 1979 1983 1987 1991 1995 1999 2003 2007 2011
Year First Elected
% MRA Dedicated to Legislative Staff
and after Gingrich’s speakership do appear to have differed from members first elected to earlier Congresses
in terms of legislative investment. Coincidence is not causation, of course. e 1995 Republican takeover
merely coincides with the beginning of a long-term, bipartisan divestment trend.
Legislative Investment Among Freshman Legislators
Despite the consistency of our findings, the mechanics of this divestment remain unclear: do these findings
primarily represent changes in resource allocation preferences among incumbent members of Congress, or
are do they result from the election of a new kind of lawmaker who has never experienced a Congress
primarily oriented toward making laws? In other words, have a new crop of electoral “showhorses” replaced
dyed-in-the-wool legislative “workhorses” (Hall 1996)? We test a series of models on freshman members
of Congress alone to explore whether institution-wide divestment in legislative staffing is driven by newly
elected members, shown in Table 3.
Models 11 and 13 mirror Models 1 and 3, testing the direct association between the year in which
members were first elected and their investment in legislative staff. Model 11 tests this association using a
dichotomous variable for whether freshmen were first elected after the Contract with America, while Model
13 uses a nominal variable to account for a linear trend in the year of first election. Estimates from Model 11
indicate that freshman members entering Congress after the Contract with America spent, on average, 5.92
percentage points less of their MRAs on legislative staff than freshman from the 103rd Congress.24 At the
mean MRA value during our period of observation ($1,157,460.00), the substantive difference in allocation
of $68,603.35 is enough money for one senior legislative staffer or two junior staffers. Model 13 indicates
that for each additional 2-year election cycle after that freshman is elected is associated with 1.07 percentage
points less of their MRAs on legislative staff spending, which is $12,426 at the average MRA level.
Models 12 and 14 examine whether the main effects of post-Contract entrance and year of first election
differ, depending on the party of the freshman member. Results from both models are consistent with
the results from Models 11 and 13. However, the interactions with party are not significant, as Figure 7
shows. e significant and substantively large negative relationship between when freshman members are
elected and the share of MRAs spent on legislative staff, coupled with the insignificant interaction with party,
again lend support for the symmetrical party competition hypothesis. e robustness of these relationships
is particularly notable given that freshmen legislators generally have not fully settled on a specific ”home
style” or approach to lawmaking. For example, given that most freshmen worry about electoral vulnerability
following their first term in office, vote share is not associated with legislative investment as it was in Tables
1 and 2. In spite of this behavioral ”noise,” recently elected freshmen clearly have invested less in legislative
24e model produces estimates of 27.1 percent vs. 21.2 percent spent on legislative staffers with all other variables held at the
Table 3: Legislative Investment Among Freshmen, Before and After Contract with America
Percentage of MRA Spent on Legislative Staff
Model 11 Model 12 Model 13 Model 14 Model 15
Elected Post-Contract
0.325∗∗ 0.220∗∗
(0.048) (0.072)
Year of Election
0.031∗∗ 0.033∗∗
(0.002) (0.004)
0.1060.041 0.027 10.607 0.007
(0.045) (0.087) (0.042) (10.028) (0.043)
Post Contract
Year of Election
Committee Chair
0.184 0.205 0.369 0.375 0.387
(0.489) (0.488) (0.458) (0.458) (0.443)
Subcommittee Chair
0.108 0.113 0.021 0.028 0.053
(0.082) (0.082) (0.078) (0.078) (0.077)
Power Committee
0.057 0.064 0.081 0.082 0.076
(0.079) (0.079) (0.074) (0.074) (0.072)
0.426∗∗ 0.484∗∗ 0.160 0.110 0.135
(0.108) (0.111) (0.104) (0.115) (0.102)
0.083+0.088+0.062 0.064 0.051
(0.047) (0.047) (0.044) (0.044) (0.043)
Vote Share
0.003 0.004 0.00005 0.001 0.001
(0.015) (0.015) (0.014) (0.014) (0.014)
Vote Share
20.00004 0.0001 0.00002 0.00002 0.00001
(0.0001) (0.0001) (0.0001) (0.0001) (0.0001)
104th Congress 0.049
105th Congress 0.088
106th Congress 0.651∗∗
107th Congress 0.149
108th Congress 0.201∗∗
110th Congress 0.423∗∗
111th Congress 0.466∗∗
112th Congress 0.593∗∗
113th Congress 0.597∗∗
Constant 0.775 0.783 60.340∗∗ 65.745∗∗ 1.080
(0.498) (0.497) (4.798) (7.044) (0.453)
N 721 721 721 721 721
R-squared 0.085 0.089 0.196 0.197 0.251
Log Likelihood 834.834 836.777 888.716 889.270 920.058
∗∗p<.01; p<.05; +p<.1
0 1
Elected Before/After Contract with America
% MRA Dedicated to Legislative Staff
Rep Dem
(a) Binary Independent Variable: Model 12
1995 2000 2005 2010
Year First Elected
% MRA Dedicated to Legislative Staff
Rep Dem
(b) Continuous Independent Variable: Model 14
Fig 7: Interactions between Member Party and Recent Election among freshman members
endeavors than have previous freshmen.
We estimate one final model on freshman members using the Congress fixed effects in Model 15. If the
ideological sabotage hypothesis is correct, we should observe the Democratic Party reversing the trend when
they regain the House majority in the 2006 wave election. If true, then incoming freshman members in
the 110th Congress should have invested significantly more in legislative staff than incoming members in
previous Congresses. is is evidently not the case. e Congress effects shown in Table 3 show negative and
significant differences between the 103rd Congress and all Congresses after the 105th. Freshman entering
the 110th and 111th Congresses led by Speaker Nancy Pelosi are not meaningfully different from those that
came before or after.
Our assessment of legislative enterprises in the era of the perpetual campaign has yielded new insights for
both political scientists and political reformers. e evidence of a post-Contract with America decline in
members’ legislative capacity sheds light on a troubling and unappreciated institutional redesign in Congress
in the last quarter century, wherein members have individually divested in legislative capacity in response
to a centralized battle for majority control of Congress. We suggest that the centralization of legislative
work within leadership offices has reduced members’ incentives to invest in their own legislative operations.
Instead, they hire legislative staffers with more generalizable skills. Doing so avoids paying premiums to retain
more expert and domain-specialized staff. What has followed is widespread member level re-allocation of
resources away from legislative staff, and towards staff that serve representation goals.25 is decline implies
that the legislators’ effectiveness also suffers, as corroborated by Crosson et al. (forthcoming).
Nevertheless, the results suggest that simply increasing member budgets and adding more bodies to con-
gressional offices will not address the broader trend of legislative divestment. Instead, it is likely to exacerbate
it, since members may shift resources towards communications and constituent service staff. Members con-
tinued to divest in legislative staff during both periods of MRA increases and Tea Party-driven budget cuts.
Moreover, members more recently elected to Congress exhibit far less commitment to legislative operations
than their predecessors, suggesting that such members face broad institutional incentives that are signifi-
cantly different from previous eras. It may also be that such legislators themselves are simply less committed
to legislating, which generates a feedback loop of legislative divestment. Individual members are unlikely to
employ and pay to retain highly skilled, policy-oriented staffers that may be in greater demand in the Senate
or on K Street unless the institution provides opportunities for legislative entrepreneurship (?). As long as
actual legislating remains so centralized, there seems little reason to pay for seasoned legislative experts when
an ambitious, and inexpensive, recent college graduate will suffice.
e general patterns we uncover here are substantiated in an interview with a Legislative Director for
a Democrat, who serves as a senior staffer responsible for the office’s legislative matters and manages three
junior colleagues. e staffer observed that the decline in legislative resources we document here is associated
with increased turnover and decreased policy experience in the House:
25See supplemental analysis in Appendix F.
e other piece is no one really knows how to legislate anymore. As someone who’s been around
a little while...there’s not really the expertise, even among committee staff that there was a couple
years ago because we’re not doing it [legislating]...I was talking to a [committee legislative staff
friend] last year. She was telling me that she had to bring all of [the legislative assistants working
for members on] the subcommittee in to teach them how to do a markup because not one of
the [legislative assistants] of this subcommittee had ever staffed their boss in a markup before.
We’ve got a great, smart team ... but two of my legislative assistants are 23 and 24...You need a
little bit more experience in a place like this. [Congress] is always gonna be a young place, it’s
always gonna attract young people that are ambitious. But it’s gone to an extreme...You can tell
that people haven’t gone through this before sometimes. Sometimes you can use that to your
advantage, sometimes it’s frustrating. But that’s a real issue...It’s hard when there aren’t people
that have been through quite a few markups, or been through a couple reauthorizations of a
major bill.
ese observations about the House are further substantiated in direct examinations of MRA expendi-
tures on legislative staff and an office staff’s experience. As summarized in Appendix E, member investments
in legislative staff do help a member to retain a more experienced staff overall. However, as such investments
have gone down, so too has experience among staffers.
For most observers of Congress, the most recognizable, systematic change in Congress over the past two
to three decades has been the rise of ideological polarization between the parties. Polarization is undoubtedly
crucial to our understanding of modern American politics: the hyper-partisanship and legislative gridlock
associated with ideological polarization impedes Congress’s ability to make laws and oversee the Executive,
especially when key democratic principles and institutional norms may be threatened. But our evidence
suggests that ideology and polarization are not the only forces systematically deteriorating members’ pro-
ductivity and legislative effectiveness. In their prosecution of the perpetual campaign to control Congress,
members have incrementally abdicated their capacity to legislate. Indeed, congressional leadership of both
parties have disincentivised their members from investing in the requisite staff to actively participate in the
legislative process.
If democratic theory suggests partisan differentiation and competition improves citizens’ ability to hold
political elites accountable, then the bipartisan decline in members’ capacity to legislate is worrisome. A
legislature supported by young, inexperienced aides—however enthusiastic and capable—is likely to fail.
e consequence is true, regardless if one’s objective is to move policy in a conservative or a liberal direction.
If both parties are to be blamed equally, and both are less able today to carry out the agendas that define their
brands, citizens have little more than carefully tailored campaign messages to retroactively assess the parties
on Election Day.
ere is room for optimism, though. If the decline in members’ capacity was created by both parties,
and impacts their policy goals equally, then it is conceivable to convince backbenchers in both parties that
improving their own legislative operations will benefit them. Moreover, doing so may counteract members
in both parties feeling frustrated by party leaders who appear to dictate the legislative process. e MRA, as
it currently functions, gives members a great deal of leeway to allocate resources as they see fit. Rather than
waiting for the rare window of opportunity to reform party and chamber institutional rules to counteract
party leaders’ and committee chairs’ extraordinary agenda setting powers, members—especially those in
relatively safe electoral positions—can rebuild their legislative enterprises to enhance their own legislative
effectiveness and institutional influence. ey need not wait for permission from the Speaker or Minority
Leader of either party to do so. Of course, doing so may run counter to the political context that caused
the decline to begin with, so we are not under the false illusion that incrementally expanding personal office
legislative enterprises in the House will solve Congress’s ills. But perhaps members who have strong legislative
enterprises will make leaders more responsive to the policy interests among the rank-and-file.
We believe that political reformers from across the ideological spectrum should welcome our discovery
that the House’s failure to fully resource its legislative operations is a bipartisan dilemma. e common
assumption that the Contract with America ushered in the “absolute frigging [sic] disaster” of a legislative
capacity crisis seemingly absolves Democrats of blame. Likewise, Republicans assume they can be credited
with some good government accomplishment. Neither of these assumptions are true. Since Democrats and
Republicans are equally to blame—and equally harmed—for the decline in member-level legislative capacity,
then they ought to be equally responsible for resolving it.
Additionally, we recognize that our observation is limited to the House, and concentrated exclusively on
member’s personal offices. We have offered limited evidence here that the legislative slack has most likely
not been picked up by committees. Further, there seems to be little evidence to suggest that members have
outsourced these responsibilities to legislative support agencies or legislative service organizations, which have
also witnessed declining budgets. Admittedly, our evidence cannot speak at all to whether Senators or Senate
committees have made the same kinds of budget allocation decisions as their House counterparts. Future
work ought to expand the scope of the analysis to include these additional institutions alongside House
member offices. Additionally, future work that adopts a more granular assessment of legislative resources—
as we do here—across a sufficiently long time to observe more than a handful of recent Congresses will offer
valuable insight into legislative enterprises.
For scholars seeking to understand how Congress changes over time, this study offers an innovative way to
comprehend the causes and consequences of legislators’ revealed preferences over their behavioral priorities.
Scholarly debates over how ideology and party affect legislative deliberation and political representation have
tended to focus on roll call votes, which, although highly important, are necessarily constrained by negative
and positive agenda powers. We show here that legislative scholars can use MRAs expenditures to observe
strategic investments that are common among all members, and are therefore not so constrained by party
leaders’ agendas. Just as the increasing ideological homogeneity within parties and growing gap between the
parties has deeply influenced Congress’s policymaking activities and outputs, we believe that the sometimes
subtle and incremental changes to Congress’s institutional make-up uncovered here have generated a legisla-
ture with important differences in power dynamics and capacity for policymaking, compared to Congresses
of previous decades.
In sum, although the House may never have taken a vote to reduce its legislative operations—and
has for the most part routinely appropriated
funding to the Legislative Branch during our period of
observation—the reductions nevertheless occurred, particularly among incoming members of both parties.
Our findings indicate that, despite the rhetoric from conservatives, the Contract did not unilaterally sabotage
Congress’s ability to function as a lean, efficient lawmaking institution. Rather, intense party competition
has encouraged members themselves to invest less in their own legislative endeavors. In response to a new
political environment consisting of nationalized elections, highly centralized policymaking processes, and
increased competition for majority status, lawmakers have incrementally dispensed with legislative staff. At
the same time, members have delegated more power to House party leaders, who responded by increasing
their own staffing resources over 200 percent (Lee 2016, Figure 5-2, 114).26 ese related developments—
centralized party leader management of the legislative process and reduced legislative operations by mem-
bers of Congress in both parties—have stunted the ability of individual policy entrepreneurs to affect policy
outputs within the chamber. As a result, we posit that highly skilled, long-serving legislative staffers are
largely over-qualified for the more routine policy-monitoring work that occurs in members’ personal offices.
Members have chosen instead to maintain smaller policy shops with higher turnover, and to allocate more
resources towards representational staff. ese changes equip members to operate within the legislature to
which they have been socialized—a legislature that values enterprises that engage in partisan warfare more
than those that engage in effective lawmaking and meaningful oversight. Without some critical institutional
intervention to disrupt this borader trend, future generations of legislators and their staff will continue to
preside over a Congress designed to win tomorrows election, and not necessarily one designed to govern
26Reproduced with the author’s permission in Appendix D, page 10 of the Supplemental Information.
Aldrich, J. (2011),
Why Parties? A Second Look
, University of Chicago Press.
Bawn, K., Cohen, M., Karol, D., Masket, S., Noel, H. & Zaller, J. (2012), ‘A theory of political parties:
Groups, policy demands and nominations in american politics’,
Perspectives on Politics
10(3), 571–597.
Bernhard, W. & Sulkin, T. (2018),
Legislative Style
, University of Chicago Press.
Brudnick, I. A. (2019), ‘Members’ representational allowances: History and usage’,
Congressional Research
Butler, D. M. (2014),
Representing the advantaged: How politicians reinforce inequality
, Cambridge Uni-
versity Press.
Cain, B., Ferejohn, J. & Fiorina, M. (1987),
e personal vote: Constituency service and electoral indepen-
, Harvard University Press.
Cox, G. W. & McCubbins, M. D. (1993),
Legislative leviathan: Party government in the House
, Cambridge
University Press.
Cox, G. W. & McCubbins, M. D. (2005),
Setting the agenda: Responsible party government in the US
House of Representatives
, Cambridge University Press.
Crosson, J., Lorenz, G., Volden, C. & Wiseman, A. (forthcoming), How experienced legislative staff con-
tribute to effective lawmaking,
T. M. LaPira, L. Drutman & K. Kosar, eds, ‘Congress Overwhelmed:
the Decline in Congressional Capacity and Prospects for Reform’, University of Chicago Press, Chicago.
Curry, J. M. (2015),
Legislating in the Dark: Information and Power in the House of Representatives
University of Chicago Press.
Fenno, R. F. (1973),
Congressmen in committees
, Little, Brown.
Fenno, R. F. (1978),
Home Style: House Members in eir Districts
, Little, Brown.
Foster-Molina, E. (2018), ‘Representing the under-privileged: Trade-offs between constituent service and
Dissertation, University of Rochester
Grimmer, J. (2013),
Representational style in Congress: What legislators say and why it matters
, Cambridge
University Press.
Hall, R. L. (1996),
Participation in congress
, Yale Univ Press.
Harden, J. J. (2013), ‘Multidimensional responsiveness: the determinants of legislators’ representational
Legislative Studies Quarterly
38(2), 155–184.
Heberlig, E., Hetherington, M. & Larson, B. (2006), ‘e price of leadership: Campaign money and the
polarization of congressional parties’,
e Journal of Politics
68(4), 992–1005.
Heberlig, E. S. & Larson, B. A. (2007), ‘Party fundraising, descriptive representation, and the battle for
majority control: shifting leadership appointment strategies in the us house of representatives, 1990–
Social Science Quarterly
88(2), 404–421.
Hertel-Fernandez, A., Mildenberger, M. & Stokes, L. C. (2019), ‘Legislative staff and representation in
American Political Science Review
113(1), 1–18.
Hibbing, J. R. (1991), ‘Contours of the modern congressional career’,
American Political Science Review
85(2), 405–428.
Kanthak, K. (2007), ‘Crystal elephants and committee chairs: Campaign contributions and leadership races
in the us house of representatives’,
American Politics Research
35(3), 389–406.
Koger, G. & Lebo, M. J. (2017),
Strategic Party Government: Why Winning Trumps Ideology
, University
of Chicago Press.
LaPira, T. M., Drutman, L. & Kosar, K. R. (2020),
Congress Overwhelmed: e Decline in Congressional
Capacity and Prospects for Reform
, University of Chicago Press.
LaPira, T. M. & omas, H. F. (2017),
Revolving door lobbying: Public service, private influence, and the
unequal representation of interests
, University Press of Kansas.
Lee, F. E. (2016),
Insecure majorities: Congress and the perpetual campaign
, University of Chicago Press.
Mayhew, D. R. (1974),
Congress: e electoral connection
, Vol. 26, Yale University Press.
McCarty, N., Poole, K. T. & Rosenthal, H. (2016),
Polarized America: e dance of ideology and unequal
, mit Press.
McCrain, J. (2018), ‘Revolving door lobbyists and the value of congressional staff connections’,
e Journal
of Politics
Miler, K. C. (2010),
Constituency representation in congress: the view from capitol hill
, Cambridge Uni-
versity Press.
Montgomery, J. M. & Nyhan, B. (2017), ‘e effects of congressional staff networks in the us house of
e Journal of Politics
79(3), 745–761.
Patterson, S. C. (1970), ‘e professional staffs of congressional committees’,
Administrative Science Quar-
pp. 22–37.
Personal communication with House House Committee on Administration staff
(2019), Author’s Personal
Petersen, E. R. (2011), ‘Congressional staff: Duties and functions of selected positions’,
Congressional Re-
search Service
Polsby, N. W. (1968), ‘e institutionalization of the us house of representatives’,
American political science
62(1), 144–168.
Poole, K. T. & Rosenthal, H. (2000),
Congress: A political-economic history of roll call voting
, Oxford
University Press.
Powell, E. N. (2015), Legislative consequences of fundraising influence,
American Political Science As-
sociation Annual Meeting, San Francisco, CA, September’, pp. 3–6.
Salisbury, R. H. & Shepsle, K. A. (1981
), ‘Congressional staff turnover and the ties-that-bind’,
Political Science Review
75(2), 381–396.
Salisbury, R. H. & Shepsle, K. A. (1981
), ‘Us congressman as enterprise’,
Legislative Studies Quarterly
pp. 559–576.
Sinclair, B. (1998),
Legislators, leaders, and lawmaking: e US House of Representatives in the postreform
, JHU Press.
Sinclair, B. (2011),
Unorthodox lawmaking: New legislative processes in the US Congress
, CQ Press.
eriault, S. M. (2013),
e Gingrich senators: e roots of partisan warfare in Congress
, Oxford University
Volden, C. & Wiseman, A. E. (2014),
Legislative Effectiveness in the United States Congress: e Lawmak-
, Cambridge University Press.
Volden, C., Wiseman, A. E. & Wittmer, D. E. (2013), ‘When are women more effective lawmakers than
American Journal of Political Science
57(2), 326–341.
... For a further analysis of the institutional debate on political aims and prevailing economic theory, see Acemoglu and Robinson (2002), who find that political elites may block innovation and institutional development. Further, partisan competition over institutional control and legislative functions has been found to decline legislative capacity (Crosson et al. 2020). While we do not seek to determine the degree to which changes in congressional politics since 1990 have altered the legislative process, this work is foundational to a forward-looking analysis. ...
... While we do not seek to determine the degree to which changes in congressional politics since 1990 have altered the legislative process, this work is foundational to a forward-looking analysis. As evidenced in the data collected, the number of legislative staffers has decreased as a share of member representational allowances (MRA's), from 25% to shy 20% from the 103 rd Congress to present (Crosson et al. 2020). The shift in resource allocation may be linked to a shift from the instructed-delegate representation model to one more similar to the trustee or sociological ...
Full-text available
This is the completed draft of my capstone paper for POSC 492 - Senior Seminar in Political Science at James Madison University. This draft was written throughout the Spring Semester as a reflection of what I have learned as an undergraduate student at James Madison University. This paper includes a review of the literature and a policy brief; both of which will be submitted to the James Madison Undergraduate Research Journal (JMURJ) and potentially other academic journals following revisions.
... In this supplemental role-data-informed but not data-dredging-ML models can help scholars investigate potential misspecifications that traditional parametric models may overlook. For example, a ML approach can account for non-linear effects often involved in models of legislative activity (for instance, logarithmic or quadratic effects of age, ideology, or seniority: Bowler, McElroy, and Muller, 2020;Crosson et al., 2019). Such models can also offer a unifying framework for truncated variables with many zeroes, like fundraising (Bonica, 2020) and lobbyist spending (McKay, 2020). ...
Full-text available
Machine learning models, especially ensemble and tree‐based approaches, offer great promise to legislative scholars. However, they are heavily underutilized outside of narrow applications to text and networks. We believe this is because they are difficult to interpret: while the models are extremely flexible, they have been criticized as “black box” techniques due to their difficulty in visualizing the effect of predictors on the outcome of interest. In order to make these models more useful for legislative scholars, we introduce a framework integrating machine learning models with traditional parametric approaches. We then review three interpretative plotting strategies that scholars can use to bring a substantive interpretation to their machine learning models. For each, we explain the plotting strategy, when to use it, and how to interpret it. We then put these plots in action by revisiting two recent articles from Legislative Studies Quarterly.
... A more thorough accounting of policy-planning organizations and the manner in which they coordinate is of central importance in understanding contemporary American political institutions. As legislative staff capacity has declined in members' offices (Crosson et al. 2019), policy-making in Congress has become centralized in party leadership (Curry 2015). We can understand this partisan legislating in the context of a recently prominent theory of American politics, which conceptualizes political parties as extended networks of policy demanding constituencies, organizations and interest groups (Bawn et al. 2012). ...
Full-text available
Think tanks have become central players in the political and policy ecosystem of the United States, yet the communication and coordination strategies and connections between them remain relatively unexamined. This paper begins to remedy that by using IRS 990 data to construct and analyze the largest board interlock network among policy planning organizations to date. An analysis of 277 policy research organizations located in Washington D.C. reveals high levels of resource seeking, issue-area coordination and weak evidence of ideological coordination among these substantively important organizations in American politics.
How are politicians informed and who do politicians seek information from? The role of information has been at the center for research on legislative organizations but there is a lack of systematic empirical work on the information that Congress seeks to acquire and consider. To examine the information flow between Congress and external groups, we construct the most comprehensive dataset to date on 74,082 congressional committee hearings and 755,540 witnesses spanning 1960–2018. We show descriptive patterns of how witness composition varies across time and committee and how different types of witnesses provide varying levels of analytical information. We develop theoretical expectations for why committees may invite different types of witnesses based on committee intent, interbranch relations, and congressional capacity. Our empirical evidence shows how committees’ partisan considerations can affect how much committees turn to outsiders for information and from whom they seek information.
Group discussion of legislation is a defining feature of parliaments and legislatures. However, there is little evidence that these conferences influence legislators' policy positions or the disposition of legislation. This article examines how, and to what extent, group discussion influences policy coalitions. It reports results from two field experiments in which state legislators spoke about randomly selected bills in a bipartisan caucus. Chosen bills received twice as much support from caucus members as unselected bills. Estimated effects are similar in magnitude within and across party lines. Discussion did not appear to affect bill-level outcomes such as bill content or bill passage, but power limitations make studying such policy-level outcomes infeasible. Qualitative data from the experiments suggest that reputation and reciprocity are key mechanisms underlying influence, but that partisan voters and media are obstacles to bipartisanship in a contemporary American legislature.
Funding for legislative staff represents a valuable commodity to legislators. However it is a resource distributed to separate internal organizations. Previous studies show strong correlations between a legislator’s available institutional power and the benefits of having access to more staffing resources. Therefore, all individual legislators, committee chairs, and party leaders face incentives to direct a larger share of the Senate’s budget. However, we argue the Senate approaches staff allocations for each organization by giving attention to the previous workload of the chamber and resources allocated to other offices in the same bill. Using newly collected data on staff resource allocations between 1885 and 2018, we observe that the Senate’s internal organizations do not undermine each other, and that allocation decisions benefit each area separately. Results of a time series model show that increasing allocations for a staff area actually promotes greater allocations for other areas rather than undermining them, and that changes in membership and eras shape how senators collectively choose to allocate staff resources.
Full-text available
Members of Congress seek to allocate their scarce staff resources carefully, given their multiple, sometimes competing, objectives. Using data on House members' staff allocations from 1994 to 2008, we demonstrate that legislators advance more (and more significant) legislation when they retain a more experienced legislative staff. This benefit, however, accrues mostly to committee chairs, whose institutional privileges allow them to leverage experienced staff, and to freshmen, whose inexperience can be best supplemented by experienced aides. Finally, we show that the positive marginal return on legislative staff experience occurs mainly among already-experienced staff. In contrast, the legislative staff for about three-fourths of House members are insufficiently experienced for these lawmakers to benefit from marginal increases in their staff's experience.
Legislative staff link Members of Congress and their constituents, theoretically facilitating democratic representation. Yet, little research has examined whether Congressional staff actually recognize the preferences of their Members’ constituents. Using an original survey of senior US Congressional staffers, we show that staff systematically mis-estimate constituent opinions. We then evaluate the sources of these misperceptions, using observational analyses and two survey experiments. Staffers who rely more heavily on conservative and business interest groups for policy information have more skewed perceptions of constituent opinion. Egocentric biases also shape staff perceptions. Our findings complicate assumptions that Congress represents constituent opinion, and help to explain why Congress often appears so unresponsive to ordinary citizens. We conclude that scholars should focus more closely on legislative aides as key actors in the policymaking process, both in the United States and across other advanced democracies.
Building on previous work on lobbying and relationships in Congress, I propose a theory of staff-to-staff connections as a human capital asset for Capitol Hill staff and revolving door lobbyists. Employing lobbying disclosure data matched to congressional staff employment histories, I?nd that the connections these lobbyists maintain to their former Hill coworkers primarily drive their higher relative value as lobbyists. Speci?cally, a 1 standard deviation increase in staff connections predicts an 18% increase in revenue attributed to the lobbyist during her?rst year. I also?nd that the indirect connections lobbyists maintain to legislators through knowing a staffer in a legislative of?ce are of potential greater value than a direct connection to a senator given a large enough number of connections. This article sheds additional light onto the political economy of the lobbying industry, making an important contribution to the literature on lobbying and the revolving door phenomenon. © 2018 by the Southern Political Science Association. All rights reserved.
This report provides basic information on congressional salaries and allowances. First, the report briefly summarizes the current salary of members of Congress, limits on their outside earned income and honoraria, and applicable health insurance and retirement benefits. Second, the report provides information on allowances available to Representatives and Senators to support them in their official and representational duties as members. These allowances cover official office expenses, staff, mail, and other goods and services. Third, the report lists the salaries of congressional officers and officials and salary limits for committee staff. Sources are hearings, reports, debates, and language of regular annual and supplemental legislative branch appropriations acts; the U.S. Code and U.S. Code Annotated Supplements to Title 2; latest Order of the Speaker of the House of Representatives, implementing a pay increase for House employees, effective January 1, 2009, issued January 9, 2009 (contained in 2 U.S.C. 60a-2a); latest Order of the President pro tempore, implementing a pay increase for Senate employees, effective January 1, 2010, issued January 5, 2010 contained in 2 U.S.C. 60a-1); the Members' Congressional Handbook (web edition), prepared by the Committee on House Administration, for members and staff of the U.S. House of Representatives, available at; the quarterly Statement of Disbursements of the House, compiled by the House Chief Administrative Officer, January 1, 2010, to March 31, 2010 (111th Congress, 1st session, H.Doc. 111-102, part 3 of 3); and the Office of Personnel Management for executive level pay rates to which some legislative employees are statutorily linked.
We argue that the leadership selection system, which now gives significant weight to fundraising, helps explain the continuing polarization of the congressional parties. Focusing first on elected party leadership posts, we demonstrate that members will select ideologically extreme leaders over "ideological middlemen" when extremists redistribute more money than their more centrist opponents. We then show that redistributing campaign money also helps ideologues win posts in the extended party leadership, though appointment to such posts by the top leaders (rather than by the caucus) makes the role of money and ideology more complex. Specifically, we demonstrate that top leaders, who are now ideologues themselves, reward the contributions of ideologically like-minded members more heavily than those of ideologically dissimilar members. This produces a more polarized leadership in Congress.
2018 Cambridge University Press. This article revisits Nelson Polsby's classic article The Institutionalization of the U.S. House of Representatives fifty years after its publication, to examine whether the empirical trends that Polsby identified have continued. This empirical exploration allows us to place Polsby's findings in broader historical context and to assess whether the House has continued along the institutionalization course-using metrics that quantify the degree to which the House has erected impermeable boundaries with other institutions, created a complex institution, and adopted universalistic decision-making criteria. We empirically document that careerism plateaued right at the point Polsby wrote Institutionalization, and that the extension of the careerism trend has affected Democrats more than Republicans. The House remains complex, but lateral movement between the committee and party leadership systems began to reestablish itself a decade after Institutionalization was published. Finally, the seniority system as a mechanism for selecting committee chairs-the primary measure of universalistic decision-making criteria-has been almost thoroughly demolished. Thus, most of the trends Polsby identified have moderated, but have not been overturned. We conclude by considering the larger set of interpretive issues that our empirical investigation poses.
Standard accounts of legislative behavior typically neglect the activities of professional staff, who are treated as extensions of the elected officials they serve. However, staff appear to have substantial independent effects on observed levels of legislator productivity and policy preferences. In this article, we use a novel data set of comprehensive longitudinal employment records from the US House of Representatives to estimate the effects of congressional staff on legislative behavior. Specifically, results from a series of heteroskedastic Bayesian spatial autoregressive models indicate that members of Congress who exchange important staff members across congresses are more similar in their legislative effectiveness and voting patterns than we would otherwise expect. These findings suggest that scholars should reconsider the role of staff in the legislative process. © 2017 by the Southern Political Science Association. All rights reserved.