How Committees Shape Legislative Behavior: An Examination of Interests and Institutions

Abstract

Whether institutional position or legislators’ interests drive legislative behavior is a central question in the congressional literature. This article focuses on the committee system in the House of Representatives and leverages the distinction between committee requests and assignments to shed new light on this debate. There is strong evidence that institutional position is largely responsible for increased legislative activity in relevant policy areas. However, there are limits to the institutional advantages of membership and its influence. The benefits of membership do not accrue to legislators immediately, and during their first term, legislators’ expressed interest increases their activity. By their second term, the power of the institution balances out these differences of interest. Legislators’ interests also play an important role in more costly behavior, like passing legislation. Legislators adapt to their institutional position, but they also remain true to the issues they believed deserved their attention when they first entered Congress.

Keywords

U.S. congress legislative behavior legislative institutions

Introduction

Every 2 years, newly elected members of the U.S. House of Representatives request committee positions. Some freshmen will receive their preferred seats while others will not. This simple fact is largely omitted when assessing how committees influence legislative behavior. A legislator’s committee request conveys an expressed interest in being on a committee and reveals an important difference among legislators. Not all committee members are alike; some wanted their seat on the committee while others did not request the assignment. Similarly, among legislators not on a given committee, some wanted to be on the committee but were not given a seat while others expressed no interest in the committee. This creates an interesting mismatch that allows one to examine how committee membership and expressed interest work together, and separately, to increase legislators’ policymaking activity.

I argue that legislators’ behavior is affected by both their interests in the policy area as expressed through requests to serve on a committee with jurisdiction and the institutional advantages that come with a position on a congressional committee. But these two conditions should matter in different ways. First, the information and expertise benefits attributed to committee membership are unlikely to be immediate. Rather, it should take some time for the benefits of belonging to a committee to take hold and to increase legislators’ behavior. Second, the influence of legislators’ expressed interest in a committee (whether rooted in constituency, ideology, personal experience, or some combination thereof) should be most notable when examining more costly forms of legislative behavior.

Based on committee requests and assignments for four cohorts of freshmen House members, and the legislation they introduced from 1987-2000, I examine the impact of legislators’ interests and positions on their behavior. Initially, freshmen committee members who expressed interest in serving on that committee distinguish themselves by authoring more legislation in relevant policy areas during their first term. But this distinction fades as their committee colleagues adapt to their unrequested position. In subsequent terms, mismatched members start sponsoring legislation at the same rate, but with less success than members who requested the committee and whose bills are still more likely to pass the House. These findings provide new evidence of how congressional institutions shape legislators’ behavior by promoting activity despite a lack of expressed interest. It also provides new insight into the limitations of institutions and the important effect of legislators’ own interests on their actions.

By leveraging the committee assignment process, I offer new evidence of legislators as responsive political actors who adapt their behavior in Congress (e.g., Grimmer & Powell, 2013; Miler, 2016; Sulkin, 2005). This research also contributes to the ongoing debate over committees as outliers who drive the legislative agenda in ways unrepresentative of the full chamber. Such concerns may be overstated when looking at proposed legislation as committee members generally act similarly regardless of their original interest in the committee. However, there remains reason to be concerned about committee outliers when it comes to policy outcomes because interested legislators are more successful in passing their legislation. This research improves our understanding of how interest and committees influence legislators’ behavior during the period examined, and has implications for today’s polarized Congress as well.

Committees and Legislative Behavior

There is consensus among congressional scholars that membership on a committee increases legislative activity in numerous ways. For instance, legislators who serve on a committee are more active in introducing legislation in that policy area (e.g., Grimmer & Powell, 2013; Sulkin, 2011; Woon, 2008) and this relationship persists when looking at broader measures of legislative efficiency (e.g., Volden & Wiseman, 2014). It is also the case that membership on a relevant committee increases the amount of federal dollars that legislators are able to direct to their constituents (e.g., Adler, 2002; Lee, 2003). However, the question of what explains committee members’ increased activity remains unresolved. Fenno’s (1973) classic work on members’ multiple motivations establishes the complexity of explaining legislative behavior, and it is to this task that I turn.

There are two schools of thought on the primary explanation for why committee membership increases legislative behavior. The first is that legislators’ interests drive their behavior. This approach emphasizes legislators’ concerns with reelection and constituency interest, and draws heavily on the distributive theory of legislative organization (Mayhew, 1974). Legislators serve on committees of interest to their constituents because doing so increases their chances of reelection. Committee membership, then, increases legislative activity because of the underlying electoral motivations of the legislator. There is considerable empirical support that legislators seek committee assignments that allow them to serve the perceived interests of their constituents (e.g., Adler, 2000, 2002; Fenno, 1973; Frisch & Kelly, 2006; Lawrence, Maltzman, & Wahlbeck, 2001; Shepsle, 1978; Shepsle & Weingast, 1987; Rohde, 1994). The interest-driven approach also can incorporate interests beyond the constituency such as legislators’ personal interest, their perception of prestigious committees, or a combination of reasons (e.g., Burden, 2007; Fenno, 1973; Frisch & Kelly, 2004; Smith & Deering, 1983).

A key implication of the distributive approach is that committees should be made up of interest outliers. Committee members represent districts with an interest in the policies under that committee’s jurisdiction, and legislators who do not have a particular interest in the issue area should not be on the committee. This logic gives rise to concerns that committees do not reflect the preferences of the full House, whether outliers are defined by interests, preferences, or partisanship (e.g., Adler, 2000, 2002; Adler & Lapinski, 1997; Carson, Finocchiaro, & Rohde, 2002; Groseclose, 1994a, 1994b; Hall & Grofman, 1990; Krehbiel, 1990; Londregan & Snyder, 1994; Shepsle & Weingast, 1987).

A critical limitation to this approach is the implicit assumption that all legislators want to be on the committee on which they serve. Quite simply, this is not the case. In the period examined here, one-third of committee seats filled by freshmen were filled by legislators who did not express interest in that committee. Committee members are not homogenous, and these differences should be part of our explanation of legislators’ behavior.

The second school of thought argues that committee members are more active due to the institutional advantages of committee membership. This approach draws on the informational theory of legislative organization, which posits that Congress is organized into a committee system to provide the chamber with expertise and information necessary for legislating (see Krehbiel, 1990, 1991). The informational model calls attention to the high levels of uncertainty and low levels of information on the floor, and identifies committees as a solution to this institutional problem. Committees are created by the full chamber to promote specialization and provide valuable information to the House. The composition of committees is driven by the collective interests of the chamber, which values the efficient provision of information. Committee members, therefore, tend to be “low cost” specialists rather than outliers (see also Gilligan & Krehbiel, 1987, 1989).

This model of legislative organization has implications for individual legislators on committees who enjoy access to specialized information and greater resources, which in turn reduce the costs of legislative activity (e.g., Ainsworth & Akins, 1997; Gilligan & Krehbiel, 1987, 1989). In contrast, legislators who do not serve on the relevant committee must bear these costs themselves. The implication is that committee members will be more active, which is consistent with the extensive empirical evidence (e.g., Hall, 1987, 1996; Lee, 2003; Woon, 2008). However, studies of legislative activity within committees also find that some committee members are more active than others (e.g., Gamble, 2007; Hall, 1987, 1996; Hall & Wayman, 1990; Perkins, 1980). This suggests that committee membership alone does not fully explain the positive relationship between committees and legislative activity.

I argue there is room here for both approaches, but as long as interests and position are conflated, it is difficult to unpack the two. Distinguishing between committee requests and membership provides a better way to understand how legislators adjust to committee membership and the extent to which internal structures constrain interest-driven behavior. The advantages of information and specialization that come with committee membership cannot be switched on. Freshmen committee members take some time to adapt to the committee, and during this period of adjustment, their decisions about whether to author legislation should be more heavily influenced by their interest. In addition to the effects of time, the influence of committee membership on behavior is also likely to be more limited when looking at costly behavior such as pushing a bill through the lawmaking process. Successful passage requires legislators’ time and political capital, and draws on legislators’ interest in the policy, which may, or may not, be separate of committee membership.

The Committee Assignment Process

The conventional notion of committee membership is dichotomous—a legislator is either on a committee or not. To better understand how committees impact behavior, however, it is necessary also to consider how a legislator got onto the committee. There are two essential stages in the committee assignments process in the U.S. House of Representatives: legislators’ requests and the party’s process of allocation. Newly elected legislators first write a formal letter to their party leaders laying out their preferred committee assignments.¹ These requests are made after the legislator is elected but before the legislator is sworn into office. Freshmen legislators’ requests reflect a mix of sincere and strategic considerations, and different legislators prefer different committees. One of the most commonly noted justifications for a committee request is the importance of the committee to the constituents in their district and the legislator’s electoral fortunes.² Incoming freshmen legislators’ committee requests, therefore, are a unique and valuable source of information about their expressed preferences for some committees over others.

One might wonder whether legislators do not request committees they want but think they will not receive. Although reasonable, the chances of such strategic behavior are low in the context of this study for several reasons. First, all committees had vacancies during the period examined here, which means that no request was impossible. Second, not all incoming freshmen want the same committees, which is reflected in the fact that every single committee was requested, including those that might seem like a long shot for a freshman (e.g., Rules), less desirable (e.g., Post Office and Civil Service), or appeal to some legislators but not others (e.g., Agriculture). For the few committees that might be widely desirable, it is not clear why freshmen legislators would not at least request a seat as they are able to request more than one committee.³ Third, I use data on all of the committees requested by each freshman legislator, not just the first ranked request. Although one might imagine strategically ranking desired committees, it is hard to imagine a legislator would not list a sincerely preferred committee at all.

Next, the party’s “committee on committees” assigns committee seats to members of their party.⁴ The selection of this group of legislators varies over history, but during the years examined here, both parties selected legislators using a combination of regional allocation and selection by party leadership (see Frisch & Kelly, 2006, chapter 2).⁵ These legislators would then allocate their party’s members to standing committees, including both incoming freshmen and any transfer requests from incumbent House members. In making freshmen committee assignments, the committee faces constraints such as the limited availability of open seats on a committee, norms of regional balance, and legislators’ expertise and electoral security (see Davidson, Oleszek, Lee, & Schickler, 2013; Deering & Smith, 1997; Frisch & Kelly, 2006; Shepsle, 1978; Uslaner, 1974). Requests are a central factor in freshmen legislators’ committee assignments because the committee has less information about incoming freshmen than their incumbent colleagues. For instance, although party loyalty may be a consideration in evaluating transfer requests or determining committee leadership, Cox and McCubbins (2007, p. 171) point out that this is not the case for freshmen legislators because their assignment comes before any votes.

It is possible that the committee on committees tries to optimize party goals beyond giving its members their requested committees (e.g., Cox & McCubbins, 2007; Rohde & Shepsle, 1973; Smith & Deering, 1983), but this is not likely in this research for multiple reasons. First, it is unlikely that party leaders make decisions about freshmen committee assignments with an eye to maximizing the number of bills sponsored, which is the dependent variable here. Second, the chronology of the process means that freshmen committee assignments are made before they cast a vote or establish a record of party loyalty. Finally, none of the factors one might think the committee on committees considers, such as incoming freshmen legislators’ fundraising record, their record of raising money from leadership sources, their professional experience in areas related to their requested committees, their electoral security, or their ideological extremism, are correlated with the dependent variable of bill sponsorship at a level greater than .06. Overall, then, freshmen legislators’ committee requests and assignments are uniquely valuable indicators of their expressed interests and institutional positions, respectively, and are well-suited to the task of untangling how committees affect legislative behavior.

The Mismatch of Committee Membership

To examine the roles of expressed interest and institutional position on legislators’ activity, it is necessary to conceptualize committee membership in a way that takes seriously variation in the paths to membership. Figure 1 provides a simple 2 × 2 illustration of committee membership as a function of whether a legislator requests a committee (horizontal axis) and whether he serves on that committee (vertical axis). The resulting figure illustrates that there are four types of legislators: (1) those who requested and received their desired committee seat (fulfilled requests), (2) those who did not express a preference for the committee on which they serve (unrequested members), (3) those whose request for a committee was denied (unfulfilled requests), and (4) those who neither expressed interest for a given committee nor were assigned to that committee. However, the literature commonly assumes that expressed interest and membership go together (i.e., types 1 or 4), which steers scholars away from the potential insights to be gained by mismatches of the two (i.e., types 2 or 3).

Figure 1.

Reconceptualizing committee membership.

Figure 1 illustrates two core hypotheses rooted in the existing literature on committees. First, committee members are expected to sponsor more legislation relevant to a committee than legislators who do not serve on the committee. Drawing on arguments that committee membership provides valuable informational benefits and resources to legislators that lower their costs of participation, we should expect to see differences in the level of activity of committee members as compared to nonmembers, including when taking account of their expressed interest in the issue area. In terms of Figure 1, the expectations for legislative activity levels are that 1 > 3 and 2 > 4.

Second, legislators who register their interest in a committee are expected to be more active in the relevant policy areas. The rich literature on legislators’ motivations—electoral, personal, policy—and the logic that legislative activity is costly combine in the expectation that legislators who request a committee will offer more bills as compared to legislators who did not request a seat on the committee, ceteris paribus. In terms of Figure 1, the expectation is that bill sponsorship patterns will be such that 1 > 2 and also 3 > 4.

The different types of legislators illustrated in Figure 1 also allow for theorizing about the relative importance of institutional position and expressed interest. I argue that interests and positions are both important determinants of legislative behavior, which sometimes work in tandem, but at other times may be in conflict. As a result, membership on the committee is important, but there are limitations to the power of institutions, and it is in those contexts that legislators’ interests will play a larger role.

This approach gives rise to expectations about the relative influence of interests and positions on legislators’ behavior. When both factors work together, as is the case with fulfilled requests (type 1), I hypothesize that legislators will be the most active. These individuals benefit from the informational and resource advantages that come with committee position as well as from the motivation that leads one to request a seat on a committee.

The advantages of committee membership, however, are expected to be lesser during a legislator’s first term. During this period, the freshman committee member is acclimating to the committee and has not yet acquired all the benefits of committee-based expertise and specialization. Consequently, differences in legislators’ expressed interest in the policy area should be a stronger predictor of early behavior as those interests are present from the first day in office. Therefore, I expect to see initial differences in legislative behavior based on who expressed an interest in a committee (1 > 2), but these differences should diminish over time.

Legislators’ expressed interest in a policy area is also expected to matter more when the costs of taking action are greater, such as when legislators work for passage of their bill. Successfully moving legislation through the House is a difficult task and often requires significant personal effort by legislators (e.g., Volden & Wiseman, 2014). As a result, I argue that legislators who made their interest known by requesting the committee are more likely to also commit their time and resources to advancing their legislation and winning its passage.

Additional Considerations

In addition to the primary hypotheses concerning committee requests and position, I also account for several alternate explanations for bill introduction activity. Legislators with relevant past professional experience are expected to sponsor more legislation because their familiarity and expertise in the issue reduces the costs of crafting legislation (e.g., Burden, 2007; Shepsle, 1978). Using official U.S. House biographies that include detailed information about the educational background and previous professional experiences of each legislator, each freshmen legislator’s biography is coded for past professional experience relevant to the policy domains of each committee.⁶

Other political and electoral factors also may affect how many bills a legislator introduces. Legislators’ electoral performance in their first election to the House may shape their activity level. Legislators who were elected with a larger percent of the vote may sponsor more legislation because they have more time to focus on their Washington, D.C., activities rather than campaigning in the district.

Additionally, party membership could be important in multiple ways. A majority of party members may introduce more legislation because being in the majority increases their bill’s chances of success. However, minority party legislators may introduce more legislation to make their voices heard in an institution that advantages the majority party throughout the legislative process. An indicator variable denoting membership in the majority party is included in the model.

Following a similar logic, the potential impact of legislators’ ideological preferences on bill sponsorship activity is also considered. More ideologically extreme legislators may sponsor less legislation because they may perceive that their bills will not be successful, or such members may sponsor more bills if sponsorship is a valuable means of expressing their policy preferences. Following convention, ideological extremity is measured as the absolute value of the distance between a legislators’ DW-NOMINATE first dimension score and the median House DW-NOMINATE score.⁷

Data and Model Specification

I examine committee requests, committee assignments, and bills sponsored for all freshmen members of the U.S. House of Representatives who joined during the 100th-103rd Congresses (1987-1994) as well as their behavior over the subsequent four terms (i.e., 1987-2000).⁸ Concentrating exclusively on freshmen eliminates the need to consider the role of seniority, past activity levels, and idiosyncratic factors such as histories with party leaders or colleagues. Data on committee requests for these new House members is taken from Frisch and Kelly (2006) who use extensive archival work to gather legislators’ committee requests that are not publicly available. These requests are a valuable measure of a legislator’s interests, which are notoriously difficult to measure.⁹ Additionally, given the strong evidence of the connection between the district and legislators’ committee preferences (e.g., Adler, 2002; Adler & Lapinski, 1997), committee requests can serve as observable indicators of legislators’ beliefs about the interests of their district. These committee request data are then matched with actual committee assignment data to determine which requests were fulfilled and which requests were not.¹⁰

Across the four incoming cohorts, 179 freshmen made a total of 507 committee requests. There are numerous cases in each of the four quadrants in Figure 1, which means that treating committee membership as a simple dichotomous measure misses real and important variation. Freshmen legislators’ requests spanned 21 different House committees, with the most popular committees being the Energy and Commerce Committee (56 requests); Science, Technology, and Space Committee (53 requests); Transportation and Infrastructure Committee (42 requests); and Banking Committee (40 requests).¹¹ Additionally, there were vacancies on every committee in all four congresses with the one exception of the Rules Committee in the 100th Congress, which means that freshmen were making requests for actual, available seats.¹²

The primary dependent variable is individual legislators’ activity as measured by the number of bills introduced in a given policy area during the first four terms they served in the House.¹³ Bill sponsorship is a critical step in the policymaking process as no policy change can occur without the introduction of a bill, even as other stages in the lawmaking process are increasingly circumvented (Sinclair, 2011). Although there are numerous ways to conceptualize legislative activity, the use of bill introductions has several advantages, most notably the fact that it is equally available to committee members and nonmembers alike. This is essential because selecting a measure of legislative behavior in which committee members have an institutional advantage by nature of their position (e.g., participation in hearings, mark-up activity, etc.) would bias the results in favor of committee member activity. Bill introductions also are a moderately costly form of legislative activity that involves commitment to the issue on behalf of the legislator and his office.¹⁴ Legislators have discretion in sponsoring legislation and bill introductions are not subject to gate-keeping by party leaders. Consequently, bill introductions are a common measure of legislative behavior used by scholars interested in measuring legislators’ nonvoting actions (e.g., Krutz, 2005; Rocca & Gordon, 2010; Schiller, 1995; Sulkin, 2005, 2011; Wawro, 2000; Woon, 2008). Data on bill introductions and their policy focus come from the Congressional Bills Project (Adler & Wilkerson, 2012).¹⁵

To determine whether a bill is relevant to a committee, I take the perspective of the freshman legislator who is deciding to sponsor a bill and their expectation of the committees to which it could be referred. Given the importance of precedent in the referral process and the possibility of multiple referrals, I look at the pattern of committees to which previous bills on the same topic have been referred. This approach reflects the reality that freshmen legislators do not know with certainty the committees to which their proposal will be referred when they write it. Past patterns of referral provide external validity as the coding conforms to real-world expectations (e.g., bills addressing technology issues are coded as relevant to the Science, Space, and Technology Committee and the Energy and Commerce Committee) and internal validity as they more accurately capture freshmen legislators’ perspective. Following previous research, a committee is considered to be relevant to a policy area if more than 20% of bills on that topic are referred to the committee (Sulkin, 2011, p. 80). I also replicate the analyses at two alternate levels of relevance to insure that the findings are not unique to the specification: a slightly higher threshold of 25% and a doubling of the threshold to 40%. The results are robust across all three standards of referral patterns and the alternate specifications are reported in the appendix.

The data are organized as legislator–committee pairs within a congress, which reflects the fact that in any given congress, each legislator could sponsor legislation relevant to any standing committee. The dependent variable is a count of the number of bills a freshmen legislator introduces relevant to each given committee in the House. As a result, I estimate the models with standard errors clustered by the legislator to take into account the multiple observations per legislator, and include congress dummies for all but one congress. I use a zero-inflated negative binomial count model, which models the two paths that could result in the observation of a legislator not sponsoring legislation relevant to a given committee: (1) He does not sponsor any legislation, and (2) he only sponsors legislation in other policy areas.¹⁶ The results of the model reported are robust to alternate specifications as well, including the estimation of a negative binomial count model and the inclusion of legislator fixed effects.

Empirical Findings

I begin with a series of models that estimate the impact of committee membership on legislative activity, which is the approach common in the literature. Table 1 presents three models that test the impact of committee membership, first for all members (column 1), and then separately for legislator–committee pairings where legislators requested the committee (column 2) and those where they did not (column 3). There is clear support for the expectation that committee members are more active than other legislators. The first column confirms that across all members, those who are on a committee sponsor more bills relevant to that committee than do legislators who are not on the committee. Most importantly, the influence of membership holds even when comparing only among those cases in which the legislator requested a committee (column 2) as well as only cases where the legislator did not express an interest in being on the committee (column 3). In particular, the data in column 2 show that those who actually sit on their desired committee (fulfilled requests) are significantly more active than their colleagues who also expressed interest in being on the committee but were not assigned to it (unfulfilled requests). The third column reveals that if legislators are seated on a committee they did not request (unrequested member), they nevertheless have access to the benefits of their committee position, and they sponsor more relevant legislation than other legislators (i.e., type 2 > type 4). Together the findings presented in Table 1 illustrate that institutional position has an important positive effect on legislators’ activity and that committee members—regardless of their interests—are more active in introducing legislation.

Table 1.

The Impact of Committee Membership on Relevant Bill Introductions, 1987-2000.

Variable	All legislators	Legislators who requested seat	Legislators without request
On committee	.847*** (.101)	.547*** (.146)	.739*** (.189)
Electoral safety (% vote in general election)	.003 (.006)	.003 (.008)	.002 (.006)
Previous professional experience	.259*** (.072)	.101 (.174)	.253*** (.082)
Majority party member	−.052 (.139)	.183 (.186)	−.065 (.144)
Ideological extremity	.424 (.420)	−.578 (.538)	.641 (.432)
Constant	.347 (.378)	1.49*** (.585)	−.248 (.383)
Wald Chi-squared (8) N	90.10*** 3716	37.50*** 507	32.30*** 3209

Note. Reported estimates are coefficients for the count process of the zero-inflated negative binomial model with robust standard errors clustered by legislator in parentheses. Congress dummy variables have been omitted from presentation. The binary process used to predict the probability that legislators will sponsor zero bills relevant to a given committee is estimated using a logit model with one independent variable: whether a legislator sponsored any legislation during that congress. N = 3716 is all freshmen–committee pairings; N = 507 is number of committee requests.

p < .10. **p < .05. ***p < .01 (two-tailed).

Table 2 shifts the focus to the impact of committee requests, and examines whether legislators who request a committee sponsor more legislation relevant to that desired committee. This approach offers a closer look at the role of legislators’ expressed interests in determining their behavior, and whether that effect differs depending on whether the legislator serves on the relevant committee. The first column in Table 2 evaluates all legislator–committee pairs and confirms that legislators who request a committee do sponsor more legislation on issues relevant to that committee than their peers, even when taking into account factors such as legislators’ previous professional experience and electoral vulnerability. When looking only at cases where a legislator is not on a committee (column 2), those who expressed an interest in the committee but did not get it (unfulfilled requests) are more active than legislators who are not on the committee but did not request it. One implication of this finding is that a legislator’s expressed interest in a committee is a valuable predictor of their behavior when institutional position provides no guidance. In contrast, column 3 examines only committee members (i.e., the legislator-committee pairs where the legislator is on that committee), and provides no evidence that committee members who requested their seat sponsor more relevant bills than their mismatched committee colleagues who did not request to be on the committee. Put differently, committee members who do not request their seat (unrequested member) are indistinguishable from their committee colleagues who want to be on the committee (fulfilled request). An important implication is that the institution compels otherwise disinterested legislators to propose legislation relevant to the committee and shape the legislative agenda. This suggests that it is service on the committee itself—not legislators’ expressed interest—that drives sponsorship activity on a committee.¹⁷

Table 2.

The Impact of Committee Requests on Relevant Bill Introductions, 1987-2000.

Variable	All legislators	Non-committee members	Committee members
Request committee	.523*** (.088)	.386*** (.108)	.179 (.214)
Electoral safety (% vote in general election)	.002 (.006)	.003 (.006)	−.005 (.010)
Previous professional experience	.286*** (.070)	.256*** (.076)	.255 (.265)
Majority party member	.010 (.139)	−.042 (.139)	.346 (.265)
Ideological extremity	.535 (.419)	.416 (.429)	1.44** (.590)
Constant	.266 (.385)	.299 (.367)	1.14* (.704)
Wald chi-squared (8) N	53.51*** 3716	23.14*** 3557	20.05*** 159

Note. Reported estimates are coefficients for the count process of the zero-inflated negative binomial model with robust standard errors clustered by legislator in parentheses. Congress dummy variables have been omitted from presentation. The binary process used to predict the probability that legislators will sponsor zero bills relevant to a given committee is estimated using a logit model with one independent variable: whether a legislator sponsored any legislation during that congress. N = 3716 is all freshmen-committee pairings; N = 159 is number of freshmen committee assignments.

p < .10. **p < .05. ***p < .01 (two-tailed).

Next, I examine the combined impact of committee requests and membership on legislators’ activity for all freshmen-committee pairings over time. Table 3 presents a model that estimates the effect of being one of the three types of legislators illustrated in Figure 1 (fulfilled request, unrequested member, and unfulfilled request)¹⁸ on legislators’ behavior over four time periods—their first term, first two terms, first three terms, and first four terms in office.¹⁹ As predicted, legislators who requested their seats (Type 1) consistently sponsor more legislation on issues relevant to that committee than their colleagues who are not on the committee and did not request to be. These legislators’ behavior is shaped by both the institutional advantages of the committee system and their expressed desire to serve on that committee. Table 3 also confirms that mismatched legislators who were placed on a committee they did not request introduce more bills than their colleagues with no interest and no relevant committee position. When legislators are on unrequested committees (Type 2), they use the advantages of their institutional position to offer more legislation on issues relevant to the committee. There is also evidence that freshmen with unfulfilled requests (Type 3) are slightly more active in sponsoring legislation relevant to that committee than other freshmen also not on the committee. This heightened activity reflects how expressed interests can shape legislative behavior in the absence of the benefits of institutional position.

Table 3.

The Impact of Committee Requests and Membership on Relevant Bill Introductions, 1987-2000.

Variable	Number of relevant bills introduced
Variable	First term	First two terms	First three terms	First four terms
On requested committee	1.10*** (.155)	1.00*** (.126)	1.09*** (.123)	.973*** (.134)
On unrequested committee	.443** (.225)	.726*** (.207)	.863*** (.204)	.942*** (.210)
Request unfulfilled	.211* (.124)	.235** (.109)	.352*** (.114)	.339*** (.131)
Electoral safety (% vote in general election)	.004 (.006)	.002 (.006)	.002 (.006)	.002 (.006)
Previous professional experience	.223** (.099)	.345*** (.076)	.336*** (.075)	.286*** (.085)
Majority party member	−.077 (.241)	−.115 (.207)	−.098 (.211)	.141 (.217)
Ideological extremity	−.316 (.447)	−.277 (.430)	−.068 (.466)	.179 (.505)
Constant	−.764* (.461)	.002 (.438)	.414 (.410)	.504 (.389)
Wald chi-squared (10) N	78.77*** 3716	120.53*** 3219	143.19*** 2893	94.33*** 2325

p < .10. **p < .05. ***p < .01 (two-tailed).

The relative magnitude of these three coefficients suggests that the largest impact of committees comes when legislators both express their interest for a committee and receive the seat. Using the observed-value approach (Hanmer & Kalkan, 2013), the predicted number of bills sponsored by different types of legislators can be compared.²⁰ A freshman legislator who both requests and serves on a committee is predicted to sponsor 5.2 pieces of relevant legislation. This is somewhat more than the predicted activity of legislators on committees they did not request (4.3 bills) and notably more than legislators whose committee requests were unfulfilled (2.5 bills). As a point of comparison, freshmen who neither requested nor serve on a committee are predicted to introduce 1.6 bills. These predicted values illustrate that institutions significantly influence legislative behavior while legislators’ interests also matter, although to a smaller degree.

In light of the earlier evidence that committee members behave similarly regardless of whether they requested their seat, I examine whether the positive relationships noted are distinct from one another. A Wald test of the hypothesis that the estimated coefficients for sitting on a requested committee and sitting on an unrequested committee are equal cannot be rejected except when looking only at legislators’ behavior during their first term in the House. During their freshman term, committee members who wanted to be on the committee sponsor significantly more relevant legislation than committee members who did not request the seat. However, this unique effect fades once behavior is assessed over the first two terms (or longer).

The data presented in Table 3 also show that legislators’ bill sponsorship activity under the two mismatched conditions are roughly equal during legislators’ first term in office. A Wald test reveals that the hypothesis that the estimated coefficients for serving on an unrequested committee and having an unfulfilled request are equal cannot be rejected. However, over a longer time period of two or more terms, the Wald test indicates that the mismatch of committee position without expressed interest has a significantly distinct—and larger—influence on legislative activity than having expressed interest alone. Together, these dynamics suggest that both expressed interest and committee position increase the number of bills that legislators sponsor, but their relative importance differs over time as freshman adjust to the House.

A Closer Look at Committee Members

Given the striking finding that committee members’ expressed interest in a committee makes no discernible difference when it comes to authoring legislation on relevant issues beyond the first term, I take a closer look at committee members separately from other legislators. Focusing only on committee members makes it possible to pinpoint the impact of an expressed desire to be on the committee while holding committee-based explanations for legislative success (such as access to more resources and information) constant. In Table 4, I estimate a model of the bills sponsored by committee members with the primary variable of interest being whether these committee members requested their seat.²¹ All other variables are included as in the previous models. Additionally, I look at different time frames to identify the longevity of the effect of expressed interest on committee-members’ activity.

Table 4.

The Impact of Requests on Relevant Bill Introductions by Committee Members, 1987-2000.

Variable	Number of relevant bills introduced
Variable	First term	First two terms	First three terms	First four terms
Requested committee	.673** (.297)	.267 (.246)	.214 (.236)	−.103 (.263)
Electoral safety (% vote in general election)	−.007 (.013)	−.007 (.012)	−.009 (.012)	−.010 (.012)
Previous professional Experience	.535** (.279)	.657*** (.247)	.461* (.263)	.520* (.308)
Majority party member	.215 (.279)	.630** (.332)	.550* (.328)	.446 (.417)
Ideological extremity	.607 (.696)	.689 (.645)	1.19** (.654)	1.14 (.841)
Constant	−.294 (.886)	.269 (.869)	.939 (.810)	1.72** (.831)
Wald Chi-squared (8) N	23.69*** 159	25.22*** 140	18.96*** 124	22.41*** 92

Note. Reported estimates are coefficients for negative binomial model with robust standard errors clustered by legislator in parentheses. Congress dummy variables have been omitted from presentation.

p < .10. **p < .05. ***p < .01 (two-tailed).

Table 4 shows that committee members’ interest in that committee initially increases their legislative activity when they are still acclimating to the institution. However, by their second term, the power of the institution appears to balance out these differences and the benefits enjoyed by all members of the committee (regardless of whether they requested their seat) level the sponsorship activity among committee members. The coefficient on the committee request variable is positive and significant at conventional levels when looking at sponsorship during legislators’ first term (Table 4, column 1), but the difference does not persist. This shift illustrates how legislators adapt their actions to reflect their institutional position. In fact, we also see some differences emerge between majority and minority party committee members after the first term, which may reflect freshmen legislators’ adjustment to their party status, and is in keeping with previous findings that majority party legislators are more active (Volden & Wiseman, 2014).²²

The impact of belonging to a committee is not instantaneous and there are delays in the benefits of institutional position as freshmen legislators acclimate. By contrast, the reasons that led them to request the committee are immediate and present. This type of behavior is consistent with recent work showing that Members of Congress are responsive to opposing campaigns or changes in district composition (e.g., Crespin, 2010; Miler, 2016; Sulkin, 2011). Notably, recent work by Grimmer and Powell (2013) finds that Members of Congress adapt their electoral behavior in reaction to losing their committee seat later in their career.

Another important finding is that although legislators on requested committees do not sponsor more bills, they are more successful in moving them through the legislative process than their committee colleagues who did not ask to be on the committee. Using data on the final status of each proposed bill from the Congressional Bills Project (Adler & Wilkerson, 2012) and the Library of Congress’s official legislative website, I examine whether bills by committee members who requested their seat are more likely to pass the House or become law.²³ The descriptive data are informative, although only suggestive given the limited number of observations. Among committee members who requested their seat, 16% of these legislators sponsor relevant legislation that successfully passes the House. This contrasts with the only 6% of committee members who did not request the committee but nevertheless sponsor successful legislation.²⁴

The data presented in Figure 2 show that 86% of successful bills authored by committee members are sponsored by members who requested their seat. Put differently, committee members who want to be there are six times more likely to sponsor a relevant bill that passes the House as compared to their committee counterparts who did not request their seat. The differential success of committee members who requested their seat and those who did not can be compared against the overall distribution of committee members (Panel 1) as well as the difference in sponsorship rates (Panel 2). Figure 2 illustrates that legislators’ interests make a key difference when looking at policy success rather than policy proposals. Committee members with an expressed interest in the committee likely care more about the issues and therefore may be more likely to invest the time, political capital, and other resources necessary to move a bill through the House (see Krutz, 2005; Wawro, 2000). In contrast, legislators who were assigned to the committee may sponsor legislation but may not have the same intensity of interest and therefore they do not do the “extra” work necessary to move their bills forward in the House. This logic also applies to the rate at which bills become law, and committee members who requested their seat again are more likely to see their bills enacted, although the overall rates are lower.

Figure 2.

The impact of committee requests.

Conclusion

Despite an extensive literature on the importance of congressional committees for legislative behavior, the question of whether this relationship is rooted in legislators’ interest in the committee or the experience of serving on the committee remains largely unanswered. Across the numerous analyses presented here, the data provide new evidence of the importance of institutional position, including information available exclusively to committee members, regardless of expressed interest. The most striking illustration of the power of the institution is the fact that after the first term, there is minimal difference in the sponsorship activity of committee members who requested their seat and those who did not. The benefits of committee membership compel even legislators who express no interest in a committee to introduce as much legislation as their committee colleagues who wanted their seats (and significantly more than similarly disinterested legislators who are not on the committee). These findings strongly suggest that the norms of committee-based work and the incentives created by the committee structure successfully promote involvement in the legislative process during this period.

It is also clear that the considerations that lead legislators to request to serve on a committee are an important driver of their behavior, particularly early in their career and when it comes to working to advance their bill through the lawmaking process. These findings reveal that there are limits to the institutional incentives that committee membership provides. The benefits of specialization do not accrue to legislators immediately and they appear to be less compelling when focusing on policy outcomes, rather than policy proposals. Additionally, legislators’ expressed interest in policy matters when they do not serve on the committee. As a result, a legislator’s request can tell us something about the legislator’s future policy activity even when they go unfulfilled.

One implication of this research is that despite some evidence of self-selection bias in committee composition, this does not necessarily translate into legislators’ behavior. What is critically revealed here is that the institution encourages committee members to be active regardless of their interest. As a result, the committee system can achieve a division of labor and the development of specialization without relying only on legislators with expressed interest. In fact, if more members sit on unrequested committees, it could serve to promote the informational and organizational advantages of the committee system for the House while mitigating the biases associated with committees composed of outliers. To the extent that the congressional agenda is shaped by the bills offered by committee members, scholarly concerns about interest-outlier bias in committees may be overstated.

Another implication of this research is that legislators’ behavior evolves in response to their institutional position, which suggests that legislators’ careers will be shaped by their committee membership regardless of their initial interest in the policy area (see also Fenno, 1973; Kellerman & Shepsle, 2009). That committees shape behavior independently is a powerful illlustration of the power of political institutions more generally. However, what is especially striking is that legislators are shaped by the structure of Congress while still being driven by their initial expressed interests. It is normatively reassuring to see that legislators remain true to the issues that they believed deserved their attention when they first came to Congress.

Finally, this research looks at a different time in Congress (1987-2000), but one not without similarities to today. Although committees played a larger role during the period examined here than they do today, party power was already on the rise. Poole and Rosenthal note that the current period of polarization began in the late 1970s (see www.voteview.com) and the years examined here include the rise of Newt Gingrich to Speaker of the House. That said, Congress today is more polarized and power is more concentrated in the hands of party leaders.

How, then, do these findings speak to the more party-driven Congress of today? A key point is that the ability of the committee system to promote activity among committee members has likely diminished because committees overall have less power in today’s party-dominated Congress. Consequently, we might expect to see more self-selection bias in legislators’ activity today as the benefits of committee membership (resources, information, policymaking role) are reduced. During the earlier period examined here, the benefits of committee membership promoted legislative activity regardless of members’ initial interest—after the first term, there was no difference in the sponsorship activity of committee members who requested their seat and those who did not. Today, however, the benefits of committee membership may not be sufficiently large to promote similar specialization and productivity. If this is the case, then legislators’ own interests (whether personal, constituency-driven, ideological, or partisan) are likely to be a stronger predictor of the bills they introduce. Although updated data on committee requests is not public, the concern would be if the smaller role of committees in recent years has resulted in a congressional agenda dominated more by individual interest outliers and less informed by collective expertise. But what this research shows is that for many years, the committee system in the House was successful in promoting issue-based expertise in the legislative process, and the potential is there for it to serve this function again in the future.

Footnotes

Appendix

Table A1.

Bill Introductions Relevant to Committee Requests, 1987-2000: Varied Time Frames and Varied Thresholds for Committee Relevance.

Variable	Number of relevant bills introduced
Variable	25% relevance threshold	40% relevance threshold
On requested committee	1.01*** (.148)	1.24*** (.182)
On unrequested committee	.805*** (.220)	.881*** (.286)
Request unfulfilled	.454*** (.125)	.649*** (.149)
Electoral safety (% vote in general election)	.004 (.007)	.003 (.008)
Previous professional Experience	.222** (.115)	.023 (.130)
Majority party member	.272 (.216)	.330 (.212)
Ideological extremity	.339 (.483)	.285 (.442)
Constant	−.057 (.473)	−.639 (.679)
Wald Chi-squared (10) N	118.45*** 3716	108.40*** 3716

Note. Dependent variable is the number of bills sponsored over first four terms. Congress dummy variables have been omitted from presentation. Reported estimates are coefficients for the count process of the zero-inflated negative binomial model with robust standard errors clustered by legislator in parentheses. The binary process used to predict the probability that legislators will sponsor zero bills relevant to a given committee is estimated using a logit model with one independent variable: whether a legislator sponsored any legislation during that congress.

p < .10. **p < .05. ***p < .01 (two-tailed).

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

Notes

Author Biography

Kristina Miler is an assistant professor in the Department of Government and Politics at the University of Maryland. Her research focuses on the U.S. Congress, political representation, and interest groups.

References

Adler

E. S.

(2000). Constituency characteristics and the “guardian” model of appropriations subcommittees, 1959-1998. American Journal of Political Science, 44, 104-114.

Adler

E. S.

(2002). Why congressional reforms fail: Reelection and the house committee system. Chicago, IL: The University of Chicago Press.

Adler

E. S.

Lapinski

(1997). Demand-side theory and congressional committee composition: A constituency characteristics approach. American Journal of Political Science, 41, 895-918.

Adler

E. S.

Wilkerson

(2012). Congressional Bills Project (1987-2000), NSF 00880066 and 0088006.1

Ainsworth

Akins

(1997). The informational role of caucuses in the U.S. congress. American Politics Quarterly, 25, 407-430.

Burden

(2007). Personal roots of representation. Princeton, NJ: Princeton University Press.

Carson

Finocchiaro

Rohde

(2002). Partisanship, consensus, and committee-floor divergence: A comparison of member behavior in the 96th and 104th congresses. American Politics Research, 30, 3-33.

Cox

McCubbins

(2007). Legislative leviathan. New York, NY: Cambridge University Press.

Crespin

(2010). Serving two masters: Redistricting and voting in the U.S. house of representatives. Political Research Quarterly, 63, 850-859.

10.

Davidson

Oleszek

Lee

Schickler

(2013). Congress and its members (13th ed.). Washington, DC: CQ Press.

11.

Deering

C. J.

Smith

S. S.

(1997). Committees in congress. Washington, DC: CQ Press.

12.

Fenno

(1973). Congressmen in committees. Boston, MA: Little, Brown.

13.

Frisch

Kelly

(2004). Self-selection reconsidered. Political Research Quarterly, 57, 325-336.

14.

Frisch

Kelly

(2006). Committee assignment politics in the U.S. house of representatives. Oklahoma City: University of Oklahoma Press.

15.

Gamble

(2007). Black political representation: An examination of legislative activity within U.S. house committees. Legislative Studies Quarterly, 32, 421-447.

16.

Gilligan

Krehbiel

(1987). Collective decision-making and standing committees: An informational rationale for restrictive amendment procedures. The Journal of Law, Economics, & Organization, 3, 287-335.

17.

Gilligan

Krehbiel

(1989). Asymmetric information and legislative rules with a heterogeneous committee. American Journal of Political Science, 33, 459-490.

18.

Grimmer

Powell

E. N.

(2013). Congressmen in exile: The politics and consequences of involuntary committee removal. Journal of Politics, 75, 907-920.

19.

Groseclose

(1994a). Testing committee composition hypotheses for the U.S. congress. Journal of Politics, 56, 440-458.

20.

Groseclose

(1994b). The committee outlier debate: A review and a reexamination of some of the evidence. Public Choice, 80, 265-273.

21.

Hall

R. L.

(1987). Participation and purpose in committee decision making. American Political Science Review, 81, 105-128.

22.

Hall

R. L.

(1996). Participation in congress. New Haven, CT: Yale University Press.

23.

Hall

R. L.

Grofman

(1990). The committee assignment process and the conditional nature of committee bias. American Political Science Review, 84, 1149-1166.

24.

Hall

R. L.

Wayman

F. W.

(1990). Buying time: Moneyed interests and the mobilization of bias in congressional committees. American Political Science Review, 84, 787-820.

25.

Hanmer

Kalkan

K. O.

(2013). Behind the curve: Clarifying the best approach to calculating predicted probabilities and marginal effects from limited dependent variable models. American Journal of Political Science, 57, 263-277.

26.

Kellerman

Shepsle

(2009). Congressional careers, committee assignments, and seniority randomization in the U.S. house of representatives. Quarterly Journal of Political Science, 4, 87-101.

27.

Krehbiel

(1990). Are congressional committees composed of preference outliers? American Political Science Review, 84, 149-163.

28.

Krehbiel

(1991). Information and legislative organization. Ann Arbor: University of Michigan Press.

29.

Krutz

(2005). Issues and institutions: “Winnowing” in the U.S. congress. American Journal of Political Science, 49, 313-326.

30.

Lawrence

Maltzman

Wahlbeck

(2001). Dispensing Opportunity and Power: The Politics of Committee Assignments in the Cannon Houses. American Journal of Political Science, 45, 551-562.

31.

Lee

(2003). Geographic politics in the U.S. house of representatives: Coalition building and distribution of benefits. American Journal of Political Science, 47, 713-727.

32.

Londregan

Snyder

(1994). Comparing committee and floor preferences. Legislative Studies Quarterly, 19, 233-266.

33.

Marrero

(2010, November 19). Nugent orients to life as freshman congressman. Tampa Bay Times. Retrieved from http://www.tampabay.com/news/politics/nugent-orients-to-life-as-freshman-congressman/1135340

34.

Mayhew

(1974). Congress: The electoral connection. New Haven, CT: Yale University Press.

35.

Miler

K. C.

(2016). Legislative responsiveness to constituency change. American Politics Research, 44, 794-815.

36.

Perkins

(1980). Influences of members’ goals on their committee behavior: The U.S. House Judiciary Committee. Legislative Studies Quarterly, 5, 373-392.

37.

Rocca

Gordon

(2010). The position-taking value of bill sponsorship in congress. Political Research Quarterly, 63, 387-397.

38.

Rohde

(1994). Parties and committees in the house: Member motivations, issues, and institutional arrangements. Legislative Studies Quarterly, 19, 341-359.

39.

Rohde

Shepsle

(1973). Democratic committee assignments in the house of representatives. American Political Science Review, 67, 889-905.

40.

Schiller

(1995). Senators as political entrepreneurs: Using bill sponsorship to shape the legislative agenda. American Journal of Political Science, 39, 186-203.

41.

Shepsle

(1978). The giant jigsaw puzzle: Democratic committee assignments in the modern house. Chicago, IL: The University of Chicago Press.

42.

Shepsle

Weingast

(1987). The institutional foundations of committee power. American Political Science Review, 81, 85-104.

43.

Sinclair

(2011). Unorthodox lawmaking: New legislative processes in the US congress. Washington, DC: CQ Press.

44.

Smith

S. S.

Deering

C. J.

(1983). Changing motives for committee preferences of new members of the U.S. house. Legislative Studies Quarterly, 8, 271-281.

45.

Sprague

(2008). The effects of measurement and methods decisions on committee preference outlier results. Political Research Quarterly, 61, 309-318.

46.

Sulkin

(2005). Issue politics in congress. New York, NY: Cambridge University Press.

47.

Sulkin

(2011). The legislative legacy of congressional campaigns. New York, NY: Cambridge University Press.

48.

Uslaner

(1974). Congressional committee assignments: Alternate models for behavior. Beverly Hills, CA: Sage.

49.

Volden

Wiseman

(2014). Legislative effectiveness in the United States congress: The lawmakers. New York, NY: Cambridge University Press.

50.

Wawro

(2000). Legislative entrepreneurship in the U.S. house of representatives. Ann Arbor: University of Michigan Press.

51.

Woon

(2008). Bill sponsorship in congress: The moderating effect of agenda positions on legislative proposals. Journal of Politics, 70, 201-216.