Committee Outliers in State Legislatures

Transcription

1 Committee Outliers in State Legislatures L. Marvin Overby University of Mississippi Thomas A. Kazee University of the South and David Prince University of Kentucky Paper prepared for delivery at the annual meeting of the Midwest Political Science Association, Chicago, IL, April 27-30, Please direct all correspondence to: L. Marvin Overby Department of Political Science University of Mississippi University, MS Voice: FAX: Abstract In this paper we extend our own previous work (Overby and Kazee 2000) exploring the prevalence of outlying committees in American state legislatures. Using an expanded sample of 44 states and measures of legislator preferences generated by a single, federated group, we find that most legislative committees are representative of the parent chambers from which their members are selected.. We discuss how this finding contributes to the generalizability of informational models of legislative organization developed to account for the U. S. Congress.

2 Committee Outliers in State Legislatures For well over a century, political scientists have recognized the importance of legislative committees. From Woodrow Wilson s (1885) dictum that congressional government was government by standing committee, to Nelson Polsby s testimony that [a]ny proposal that weakens the capabilities of congressional committees weakens the Congress (Hearings on Committee Organization..., 1973), to Davidson and Oleszek s (1998) observation that committees are the centers of policy making, oversight of federal agencies, and public education (emphasis added), it has been the received wisdom that committees are vital to the operations of the United States Congress. There is far less consensus, however, on exactly why legislative committees are so important... or, indeed, what their primary purpose is. The three theories most commonly proposed to explain committee purpose and power --- the distributive perspective, the party-dominant perspective, and the informational (or chamber-dominant) perspective (for a succinct overview, see Deering and Smith 1997) --- posit conflicting claims regarding not only what core functions committees perform, but also regarding the fundamental organizational principles of the Congress. What is more, at their stylized extremes, these theories make assertions that go far beyond the Congress to draw much more general conclusions concerning how any rational legislature (Gilligan and Krehbiel 1990) should be organized. In this paper, we attempt to shed additional empirical light on these claims by extending our own previous work examining legislative committee compositions in the American states. Using data from a 44-state sample, we provide evidence that committee compositions are broadly representative of the full chambers from which their members are drawn, further buttressing the informational theory of legislative organization. 1

3 Theories of Legislative Organization Theories of legislative organization have recently received exhaustive scholarly attention and require only brief recapitulation here. For years, the textbook portrayal of the Congress held that it was a highly distributive institution and that its members had created a largely autonomous committee structure to reflect this ethos (Wilson 1885; McConachie 1898; Fenno 1966, 1973; Manley 1970; Mayhew 1974; Price 1978; Davidson 1981; Weingast and Marshall 1988). In such a decentralized institution, members are more or less free to self select onto committees of special interest to their constituents (Shepsle 1978), and this in turn permits them to exert disproportionate influence over policies that will help ensure their reelection. Since legislators do not perceive of themselves as engaging in a zero-sum transactions (Mayhew 1974), they regularly logroll with each other, trading for influence in areas of interest to them in exchange for deference to others on matters on which they care little. While this might result in a chamber composed of very unrepresentative committees, it also fosters gains from trade that can lead to an efficient allocation of resources, the development of significant expertise, and an expedited policy making process (Shepsle and Weingast 1995; Deering and Smith 1997). Theorists associated with the informational perspective disagree with the distributional model on several significant points. First, they believe that the textbook, distributive view seriously underestimates the basic majoritarian nature of the Congress (and similar legislative institutions). Since bills must ultimately face the judgment of the floor and since it is the chamber that empowers committees, Krehbiel (1991) and others think it appropriate to conceive of committees as agents of the more powerful floor. Second, they hold that the chamber has substantial motivation to keep its committee agents on relatively short leads. For while a 2

4 distributional arrangement might be rational under stylized conditions of full information, such is never the case in the real world of legislative activity. In real legislatures, the full implications of legislation are rarely immediately obvious, and hence the larger chamber usually relies on its committees to research and report costly information. Such committee-chamber cooperation is more likely when committee preferences are similar to those of the parent chamber or, if nonrepresentative (i.e., outlying) committees cannot be avoided, by limiting their mischief through strict restrictions on their jurisdiction and by the utilization of subsequent rules that facilitate chamber attempts to undo committee actions (Krehbiel 1991; Gilligan and Krehbiel 1990). In sum, the informational perspective holds that collective informational motives not individual electoral ones drive legislative organization and that the larger legislative institution has both the motives and the means to construct a committee system that serves collective rather than individual ends. The party-dominant perspective shares with informational models the assumption that committees ultimately are agents responsible to a more powerful principal. In this case, though, the principal is the parties, and particularly the majority party, not the full chamber (Liebowitz and Tollison 1980; Kiewiet and McCubbins 1991; Cox and McCubbins 1993). Like the distributional model, however, the party-dominant perspective accepts that electoral concerns weigh heavily on the minds of legislators. Here, though, the logic has is that members believe their individual electoral fortunes are closely tied to their party s success in the institution... and that party leaders have access to numerous devices (e.g., influence over committee assignments, control of the agenda, access to special rules) with which to encourage cooperation (Rohde 1991). 3

5 Attempts to test these differing perspectives have produced mixed empirical findings, although overall more support for the informational and party-dominant models than for the distributional one. In an important 1990 article, Krehbiel demonstrated that in the 96 th through the 99 th Congresses only the Armed Services Committee was consistently an outlier and, following his lead, others have produced similar results (see, e.g., Groseclose 1994; Brown, et al. 1997). In the party-dominant tradition, studies by Cox and McCubbins (1993) and Kiewiet and McCubbins (1991) have also uncovered a small number of committee outliers, in particular among important control committees like Rules, Appropriations, and Ways and Means that structure chamber deliberations, produce proposals that affect essentially every member, and are, therefore, of particular interest to party leaders. Others have challenged these findings, often noting that the roll-call-based interest group measures used by Krehbiel and others bias studies toward null results because they are not well tailored to the jurisdiction-specific hypotheses being tested and because floor deference to committee members could result in nonmember voting patterns that mirror committee voting patterns, even though the committees were, in fact, biased and powerful outliers (Hall and Grofman (1990). Using constituency-based measures instead of interest group scores, Hall and Grofman (1990) find substantial evidence of committee (and subcommittee) outliers in the 99 th Congress. 1 Further support for this position has been provided by Adler and Lapinski (1997) 1 Krehbiel (1994) has, in turn, answered both of these critiques. Drawing a sharp distinction between the geographical constituencies used by Hall and Grofman and the election constituencies that legislators care far more about, he has argued that constituency-characteristic measures are inferior to vote-based measures as predictors of voting behavior on key legislative decisions. Second, Krehbiel (1994) argues that the Type II error risk from deference is exaggerated. Since the measure of true preferences is disturbed by a constant term in the pro-committee direction, and since the disturbances in the mean differences and the variance expectations cancel each other out, the t-statistics used for 4

6 who searched for outliers employing district-level measures of policy need. Using a 50-year time series, they compare 13, non-exclusive committees with the floor and find that several committees are overrepresentative of members whose districts possess qualities of particular concern to that panel. 2 Still others have argued for the use of scaled algorithms such as NOMINATE scores (Poole and Rosenthal 1991), since interest groups ratings exaggerate the degree of extremism and bipolarity in Congress (Snyder 1992), leading to artificially inflated variances in t-scores and, again, a bias toward null results. 3 Using NOMINATE scores and a times series involving all House committees in 17 post-world War II Congresses, Londregan and Snyder (1994) conclude that one-third are outliers, including --- in many instances --- even important control committees (i.e., Appropriations, Rules, and Ways and Means). In the conditional party government tradition (Rohde 1991), some recent research has argued that studies focus on the conditions that lead to either committee-, party-, or chamberdomination. Maltzman (1995), for instance, finds that while none of the 14 committees in his sample was an outlier compared to the floor and that control committees are, in fact, significant[ly] representative of the floor, majority party committee delegations often differ from their parent caucus, even on control committees. He concludes that this is compatible with inferences regarding outliers are left unbiased. See also Herron (1999) for mixed evidence regarding the effects of artificial extremism in interest group ratings. 2 Adler and Lapinski concede that the majority of outlying committees they uncover are traditionally considered to be private goods committees (e.g., Agriculture, Interior and Insular Affairs), while there are few outliers among policy committees. Notably, they do not even consider control committees, stipulating that such committees extremely unlikely to be outliers. 3 Based on simulations bounded within the range of likely observations, Krehbiel (1994) demonstrates that --- as with deference --- this concern is probably overstated, since the probability of an extremism-based Type II error is only marginally greater than that universally accepted when using customary 95 percent confidence intervals. 5

7 a conditional party-dominant model because party leaders are apt to appoint extreme delegations... [in order] to counter-balance minority party delegations, especially on vital institutional maintenance committees (p. 675). This evolving consensus around a conditional model is perhaps best encapsulated by Deering and Smith (1997) in their textbook on congressional committees: None of these perspectives [i.e., the distributive, chamber-dominant, or partydominant perspectives] full captures the nature of the relationship among committees, parties, and the parent chambers. Rather, each exaggerates the importance of certain features of congressional rules and practices. In fac, policy making through committees represents an admixture of these alternate models... In fact, we believe that the exercise of committee discretion, party influence, and chamber influence is visible in some degree on most major bills passed by Congress (pp. 4-5). Committee Outliers in State Legislatures As controversy continues along both theoretical and empirical dimensions, it is important to reiterate the point that virtually all of the evidence marshaled in the committee outlier debate has been gathered from only one legislature. Even though all camps in the debate tend to frame their arguments with universalistic rhetoric, the empirical bases for these claims come almost exclusively from studies of only one chamber of one legislature --- the U. S. House of Representatives. 4 In an earlier paper (Overby and Kazee 2000), we utilized the federal structure of the United States to gain added purchase on the outlier question. 5 Arguing that state data 4 But see studies by Marshall, Prins, and Rohde (1997) on Senate committees, and Ainsworth and Akins (1997) on congressional caucuses. 5 See also Battista (1998) who found little evidence of committee outliers in the lower houses of the Minnesota, Iowa, Connecticut, and New Hampshire state legislatures. 6

8 would allow us to test the generalizability of legislative organization theories developed in the congressional environment and that there are substantive reasons in an era of policy devolution to be concerned with state committee representativeness, we employed a sample of 12 state legislative lower chambers to search for committee outliers at the sub-national level. Across a fairly good cross-section of states 6 and using a variety of different measures of legislator preference, 7 we found very little evidence of outlying committees. Among non-control committees, rates of outliership ranged from 0 percent (found in seven states) to 21 percent (in Louisiana), with an average of 6 percent per state. We found no evidence in any state in the sample of outliers among the important appropriations, finance, and rules committees that exercise disproportionate control of legislative activities. We found somewhat higher incidents of committee partisan delegations that differed significantly from the parent caucuses, although here, too, outliers were the exception rather than the norm: 9 percent in the case of Democratic delegations, 10 percent in the case of Republican delegations. 8 In this paper, we extend this initial analysis through a substantial expansion of the sample size, from 12 states to 44. We believe there are compelling reasons to do so. First, the initial sample was largely a purposive one. Although the original sample included a good deal of 6 The states included in the sample were Arkansas, Colorado, Illinois, Kentucky, Louisiana, Mississippi, New York, North Carolina, Ohio, South Carolina, Virginia, and Washington. 7 We utilized four types of interest group measures: ratings based entirely on roll call voting behavior, modified roll call-based measures (i.e., those that either use survey results to supplement rollcall scores or add/subtract points to reward/punish legislators for efforts such as bill sponsorship), constituency characteristic-based measures, and survey-based measures. 8 Unlike Maltzman (1995), we did not find consistency in the direction of partisan delegation extremity (e.g., Democratic delegations were as likely to be more conservative as more liberal than their parent caucus). 7

9 diversity in terms of legislative professionalization, 9 party control 10, and the balance of power among committees, party leaders, 11 and party caucuses, it was not random, but rather generated by our access to various interest groups that rate state legislators. By expanding the number of cases to include virtually the entire universe of states, we can have greater confidence that our earlier results were not driven by the cases included in the non-random sample. Second, increasing the number of cases provides leverage to undertake multivariate analysis. To the extent that the occurrence of committee outliers does vary by state, multivariate analysis can help us isolate and examine the institutional factors that contribute to the emergence of nonrepresentative committees. Data and Methods In this paper, we expand our original 12-state sample by adding an additional 32 states. For each state, we have obtained ratings of members of the lower legislative chamber compiled by the state chapter of the National Federation of Independent Business (NFIB). NFIB describes itself as the largest advocacy organization representing small and independent businesses in the country, with a membership of 600,000 business owners, who employ more than seven million 9 Using Squire s (1992) index of professionalization, the states in the sample ranged from the most professionalized (New York) to the 44 th most professionalized (Kentucky), with four in the top third, six in the middle third, and two in the bottom third. 10 Seven of the 12 chambers were controlled by the Democratic party, five by the GOP. 11 At least one legislative chamber in our initial sample fell into five of the six categories of committee-leader-caucus powers identified by Francis (1989), including an over-representation of chambers (25 percent) identified by Francis as having powerful committees. 8

10 people, and a legislative office in each state capitol. 12 The six states excluded from this analysis fall into one of four categories. Maine and Connecticut use joint House-Senate committees that complicate calculating committee-chamber comparisons. Nebraska has a unique unicameral legislature that renders it analytically distinct from the other states. NFIB does not currently have ratings posted for members of the New Hampshire and New Jersey legislatures. And, finally, in Hawaii, the NFIB compiles not a rating, but merely a listing of Friends of Small Business, which makes data from that legislature incommensurable with that from the others. 13 The NFIB generally takes a conservative, pro-business stance on legislation and sometimes takes stances on issues that are not directly (or exclusively) related to small businesses. Votes included in the ratings for some of the states in our analyses included tort reform, legislative term limits, privatization of government functions, regulatory reform, limits on government spending, repeal of minimum wage laws, educational reform to stress basic skills, and a variety of tax cuts... all positions that the NFIB endorsed. A listing of the states in the original analysis and those in the current expanded sample is found in Table [Table 1 About Here] 12 Description taken from the organization s website, 13 The Hawaii branch of the NFIB justifies this by noting that the Hawaii legislature seldom allows an issue to reach the floor for a final vote unless legislative leaders are assured passage. There are few no votes on final passage. The Voting Record usually showed each Hawaii legislator with a near-perfect, 100 percent, voting record on small-business supported bills (see the entry for Hawaii). 14 In our original analysis of South Carolina (Overby and Kazee 2000), we reported two ratings of members of the South Carolina legislature calculated by the state s Business and Industry Political Education Committee. One is based on the strength of business interests in each member s district; the other is based on members roll call voting behavior. Here we use only the roll call-based rating, which is more analogous to the NFIB scores used for most of the states in this analysis. 9

11 In each of the 32 new states added in this analysis, the state chapter of the NFIB rated the members of the lower chamber of the state legislature, using legislators votes on bills of particular interest to the groups leadership. In each case, legislators are ranked according to their support for the NFIB s agenda with scores reflecting the percentage of times they agreed with the group s position on the pieces of legislation used in the rating. Higher scores reflect a higher level of support for the NFIB agenda and, more generally, conservative voting patterns. Summary statistics related to these ratings are included in Table 2. While not all of the state-bystate ratings are particularly well behaved and there is a pronounced upward bias in most scores (reflecting a general pro-business bias in the state legislatures), in most cases there is some considerable variation in the range of scores. Since different state legislatures have different agendas, and consider and vote on different pieces of legislation, no behavioral measure of member preferences can be entirely commensurable across states. However, the use of ratings generated by a federated group such as the NFIB at least ensures that in each state we are measuring more or less legislative preferences across a similar dimension and in a roughly similar manner. [Table 2 About Here] Data on members committee assignments were gathered from relevant issues of the State Yellow Book. In all cases, we examine only committees composed entirely of members from the state s lower legislative chamber; in some cases (e.g., Oregon) excluding joint committees substantially lowers the number of committees in the analysis. To test for the presence of committee outliers, we rely most heavily on difference of means tests (i.e., two-sample t-tests) to ascertain whether or not there statistically meaningful 10

12 differences between the preferences of committees and those of their relevant parent chambers. While difference in medians are more useful for purposes of formal modeling and have, therefore, become the de facto industry standard in most committee outlier research, we believe there are numerous reasons --- both statistical and substantive --- for using t-tests (e.g., their greater transparency, their greater sensitivity to outliers, the utility of means for estimating medians, our belief that in real legislative settings the intensity of preferences matter). Since we have outlined those justifications in detail elsewhere (see Overby and Kazee 2000), we will simply note here that difference of means tests and difference of medians tests (whether computed using Monte Carlo simulations or Wilcoxon rank-sum tests) yield virtually identical results (see Groseclose 1994; Londregan and Snyder 1994; and Overby and Kazee 2000). Findings We report our initial findings in Table 3, which lists the number of control and noncontrol committees by state and, in the column on the far right, records the percentage of committees in each category that are outliers compared to their parent chambers according to mean interest group ratings. 15 One finding is immediate and obvious: outlying committees are rare. Of the 115 control committees that we identify in our 44-state sample, only 3 percent (n = 4) are outliers. Similarly, of the 764 non-control committees, only 4 percent (n = 34) are outliers. In a majority of the 44 states (n = 25 or 57 percent), there are no instances of outliers of any 15 We counted as control committees those whose names indicated they were responsible for tax bills, appropriations choices, or the rules of the chamber; all others we coded as non-control. A complete list of committees of both types is available from the authors. 11

13 kind. 16 [Table 3 About Here] Table 4 identifies, by state, which committees are outliers relative to their parent chambers and in which direction (i.e., more conservative/pro-business or more liberal/antibusiness). While there are relatively few cases on which to base any sort of trends, there does seem to be at least a faint pattern in the distribution of outliers, which split almost evenly between those that are more conservative and those that are more liberal than the chamber. As a general rule, the few committees that are outliers in the conservative/pro-business direction tend to be those with jurisdictions over matters like banking, insurance, and natural resources. Such committees undoubtedly routinely deal with matters of particular interest to conservative/probusiness constituencies. Conversely, the few liberal/anti-business outliers tend to have jurisdictions over matters such as labor, cities, health, and education... issue areas where the most active constituencies are likely to be relatively liberal/anti-business. [Table 4 About Here] Of the four control committees that are outliers, three (i.e., the finance committees in West Virginia and Wisconsin, and the Rules Committee in Wyoming) are conservative/probusiness outliers, while only one (i.e., the Ways and Means Committee in Vermont) is liberal/anti-business. While the pattern of outliers seen in Table 4 comports well with a distributive perspective on legislative organization (i.e., legislators migrating to committees on which they can serve 16 As with the previous studies cited earlier, difference of medians tests yield virtually identical results, both in terms of the total number of outliers identified and the particular committees so designated. 12

14 constituency interests), the overall paucity of outliers in our sample speaks much more strongly in favor of the informational perspective. The fact that only one committee in 25 is an outlier is consistent with institutional efforts to create representative committees that will serve as loyal agents of the floor in terms of acquiring and divulging important legislative information. If NFIB scores (and the other interest group ratings we use) are read narrowly only as jurisdiction-specific indicators of support for small business, the meager number of outliers uncovered in this study can be interpreted as strong evidence that most state legislative committees that regulate, monitor, and oversee business matters in their respective states have not been captured by business s most ardent advocates within the legislature (nor, conversely, by anti-business interests). The plethora of non-outlying committees in our sample that deal with such matters as labor-management relations, commerce, industrial development, tourism, natural resources, public lands, banking, insurance, consumer protection, environmental regulation, utilities, business regulation, transportation, and the like dwarf the small number of business-related committees that do emerge as outliers. In sum, when creating committees to handle matters of the greatest relevance to small business, state legislatures generally behave as though they are motivated primarily by informational concerns, constructing panels whose mean preferences on business-related matters are quite close to those of the full chamber. If NFIB (and similar) scores can be interpreted more broadly, as surrogate measures of members general ideological inclinations, the point can be extended. Within the ideological spectrum represented by their members, the dearth of committee outliers indicates that state legislatures seek to construct committees with central philosophical tendencies that closely mirror the floor... a point that is especially important in terms of the control committees that structure key legislative allocational 13

15 decisions. We also examine here the relationship between party delegations on the committees included in this analysis and their caucuses in the larger chamber. These results are summarized in Table 5. [Table 5 About Here] The party-dominant perspective on committee organization holds that the party delegations on committees should be generally representative of the larger caucus, and that is generally what we find. Although party delegation outliers are more common than committee outliers, they are still certainly the exception rather than the rule. Among Republican committee delegations, we find that only 11 percent of control committees and eight percent of non-control committees are outliers when compared to the Republican caucus in their respective chamber. The figures are roughly the same for Democratic committee delegations, where we find eight percent of delegations on control committees and nine percent of those on non-control committees to have mean NFIB ratings that differ significantly from their respective caucus average. 17 What is perhaps more interesting is that the direction of committee delegation outliers is often not as predicted. In his work on the U.S. House, Maltzman (1995) found evidence that Democratic delegations on important control committees tended to be liberal outliers compared to the Democratic caucus. He interprets this as evidence of conditional party control of committee composition, arguing that this results from the fact that Democratic party leaders 17 These findings differ somewhat from Maltzman s (1995) analysis of the U.S. House of Representatives, where he reports party delegation outliers only on important control committees. Here we find more on less influential, non-control committees. 14

16 packed important committees with liberal Democrats, anticipating that Republicans would do just the opposite and that they would need relatively liberal membership on these committees to balance out the conservative GOP members. We find considerably more mixed evidence of that in our sample of state legislative committees. Specifically, of the 12 outlying Republican delegations on control committees, five (42 percent) are outlying in an unexpectedly liberal direction; even more telling, of the nine outlying Democratic delegations on control committees, fully seven (78 percent) are conservative outliers compared to their parent caucuses. 18 If conditional party control of committee composition is occurring at the state level, it is far less obvious than it is at the federal level. Conclusions In this paper, we have advanced the literature on legislative organization to take a more comprehensive look at state legislatures. Replicating and expanding the analysis first presented in Overby and Kazee (2000), we have examined committee compositions in the lower chambers of a sample of 44 states relying largely on legislative ratings generated by the National Federation of Independent Business. In general, our findings provide strong evidence in support of an informational model of legislative organization. The fact that only four percent of all committees and only three percent of important control committees are outliers relative to their parent 18 Unlike the practice in the U.S. House of Representatives, minority parties in state legislatures do not always have the right to name their own committee delegations (although minority party leaders are often consulted, either formally or informally). Because of this, it is important to note that for both the Republican and Democratic analyses a majority of outlying committee delegations that are in the wrong direction (i.e., three of the five Republican cases; four of the seven Democratic cases) come from state houses where that party is in the majority. 15

17 chambers indicates that, across states, legislative committees are composed in a manner that makes them broadly representative of the floor. Substantively, while this does not necessarily mean that business interests are not over-represented in state legislatures, 19 it does provide evidence that relative to the chambers conservative, pro-business interests are not overrepresented on state legislative committees... especially not on the vital institutional maintenance committees that structure the legislatures most important taxing and spending decisions, and generally not even on the committees with jurisdictions over issue areas most relevant to business. In terms of legislative theory, these findings provide robust evidence that our previous findings (Overby and Kazee 2000) were not the result of our original sample size and, therefore, that informational models of how rational legislatures are organized are generalizable beyond the United States House of Representatives. The dearth of committee outliers in our sample is highly consonant with the perspective of Krehbiel and others that committees are constructed in such a way that they will be loyal agents of the floor, so that they can be relied upon to collect and divulge the information their parent chambers need to arrive at reasonable policy decisions. While we plan further work to build upon these findings, 20 at this juncture we think it safe to paraphrase Deering and Smith (1997:109) to the effect that the preponderance of evidence available from both Washington and the state capitals indicates that at best, only a relatively small portion of legislative committees are significantly enough different to sustain the distributive committees theory. 19 Indeed, the positive bias in NFIB scores noticed in most states indicates that businesses have many friends in state legislatures, a phenomenon that is probably rooted in candidate emergence and campaign finance patterns. 20 For example, expanding the time series, using other interest group ratings and/or developing state-level NOMINATE scores, and using multivariate analyses to tease out the factors responsible for the limited variance observed in the prevalence of outliers in the states. 16

18 Table 1 States Included in the Analysis State Interest Group States included in original Overby and Kazee (2000) analysis States new to this analysis Arkansas Colorado Illinois Kentucky Louisiana Mississippi New York North Carolina Ohio South Carolina Virginia Washington Alabama Alaska Arizona California Delaware Florida Georgia Idaho Indiana Iowa Kansas Maryland Massachusetts Michigan Minnesota Missouri Montana Nevada New Mexico North Dakota Oklahoma Oregon Pennsylvania Rhode Island South Dakota Tennessee Texas Utah Vermont West Virginia NFIB a NFIB a Chamber of Commerce KY Forward LABI b BIPEC c CPNY d NCCPPR e NFIB a BIPEC c VA Free NFIB a NFIB a Wisconsin Wyoming a National Federation of Independent Business; b Louisiana Association of Business and Industry; c Business and Industry Political Education Committee; d Conservative Party of New York; e North Carolina Center for Public Policy Research 17

19 Table 2 Summary Statistics of NFIB Ratings for New States State Year/s Range µ ) State Year/s Range µ ) Alabama Montana Alaska Nevada Arizona New Mexico California North Dakota Delaware Oklahoma Florida Oregon Georgia Pennsylvania Idaho Rhode Island Indiana South Dakota Iowa Tennessee Kansas Texas Maryland Utah Massachusetts Vermont Michigan West Virginia Minnesota Wisconsin Missouri Wyoming

20 Table 3 Committee Outlier Percentages by State State Committee Type Total # of Committees % Outliers Alabama Alaska Arizona Arkansas California Colorado Delaware Florida Georgia Idaho Illinois Indiana Iowa

21 Table 3 continued Kansas Kentucky Louisiana Maryland Massachusetts Michigan Minnesota Mississippi Missouri Montana Nevada New Mexico New York North Carolina

22 Table 3 continued North Dakota Ohio Oklahoma Oregon Pennsylvania Rhode Island South Carolina South Dakota Tennessee Texas Utah Vermont Virginia Washington

23 Table 3 continued West Virginia Wisconsin Wyoming All States All Committees

24 Table 4 Outlying Committees by State State Type Committee Names State Type Committee Names Delaware None Oklahoma None Labor Mental Health; Transportation Florida None South Carolina None Agriculture Medical, Military, Public, & Municipal Georgia None Banking; Defense/Veterans; Health; Insurance; Special Rules South Dakota None Education; Health and Human Resources Idaho None Texas None Agriculture Land/Resource Management Kansas None Utah None Calendar and Printing Education; Energy Kentucky None Vermont Ways and Means Banking/Insurance Fish and Wildlife Louisiana None West Virginia Finance Labor; Municipalities; Natural Resources None Massachusetts None Wisconsin Finance Personnel None Mississippi None Wyoming Rules Conservation; Military; Oil/Gas; Utilities; Investigations; Library Judiciary; Transportation New York None Committees outlying in a liberal direction relative to the floor are italicized; committees outlying in a conservative direction relative to the floor are not. Children/Families; Social Services 23

25 Table 5 Party Delegation Outliers by State State Committee Type Republican Delegations Democratic Delegations Alabama 0 0 Judiciary; Local Government; Local Legislation 1 Local Legislation 2; Local Legislation3; Local Legislation 5 Alaska 0 0 Resources Committee on Committees Arizona 0 0 Public Institutions Environment; Government Reform Arkansas California 0 Rules Agriculture; Consumer; Elections; Water Insurance; Public Safety Colorado 0 0 Agriculture; Local Government Agriculture Delaware 0 Revenue/Finance Labor Business; Economic Development; Housing Florida 0 0 Environment; Government Operations; Regulated Services Agriculture; Educational Appropriations; Water Georgia 0 0 Defense Children; Health/Ecology; Insurance; Special Judiciary Idaho 0 0 Agriculture 0 Illinois Revenue 0 Cities; Consumer Judicial/Crime 24

26 Table 5 continued Indiana 0 0 Aged; Natural Resources; Transportation Agriculture Iowa 0 Ways and Means Commerce; Education Agriculture; Natural Resources Kansas Appropriations Appropriations Calendar; Interstate Cooperation 0 Kentucky 0 0 Judiciary Banking Louisiana Agriculture; Municipalities; Natural Resources Maryland 0 0 Environment; Judiciary Judiciary Massachusetts Ethics; Personnel Michigan 0 0 Senior Citizens Commerce; Utilities Minnesota 0 0 Regulation Education Mississippi Rules 0 Conservation; Jobs; Oil; Salaries Offices Missouri Appropriations-Natural Resources 0 Labor Ethics; Federal/State; Higher Education; Utilities Montana 0 Appropriations 0 0 Nevada

27 Table 5 continued New Mexico Rules New York 0 0 Small Business; Tourism Education; Government Operations; Local Government; Social Services; Transporation; North Carolina 0 0 Judiciary 3; Utilities; Government Operations Agriculture; Business Travel; Ethics North Dakota 0 Rules Judiciary 0 Ohio 0 0 Judicial/Crime; College/University 0 Oklahoma Rules; Appropriations 0 Government Operations Commerce; Mental Health; Public Health; Tourism; Transportation Oregon Pennsylvania Finance 0 0 Education Rhode Island South Carolina Rules 0 0 Medical; Memorials South Dakota 0 0 Health/Human Services 0 Tennessee 0 0 Conservation; Consumer 0 26

28 Table 5 continued Texas n/a 0 n/a Agriculture; Transportation; Urban Affairs Utah 0 0 Education Energy Vermont Ways and Means 0 Judiciary Housing/Military; Transportation Virginia 0 Rules Education Cheasapeake Bay; Health; Mining Washington 0 0 Financial/Insurance; Health; Law/Justice Energy/Utilities West Virginia n/a Finance; Rules n/a 0 Wisconsin Finance 0 Child/Family; Financial Institutions; Labor Assembly Organization; Insurance; Mandates Wyoming Rules 0 Transportation Judiciary All States (n) 12 9 (n) All Committees (n) Republican committee delegations in Texas and West Virginia are excluded from the analysis, since there is no variance in Republican members NFIB ratings in those states. Committee delegations outlying in a liberal direction relative to their parent caucus are italicized; delegations outlying in a conservative direction relative to their caucus are not. 27

29 Sources Adler, E. Scott and John S. Lapinski Demand-Side Theory and Congressional Committee Composition: A Constituency Characteristics Approach. American Journal of Political Science 41(July): Ainsworth, Scott H. and Frances Akins The Informational Roles of Caucuses in the U. S. Congress. American Politics Quarterly 25 (October): Battista, James S. Coleman Organizing State Legislatures: Testing Theories of Legislative Committees at the State Level. Paper presented at the annual meeting of the American Political Science Association. Brown, Robert D., Jennifer M. Davis, L. Marvin Overby, Charles E. Smith, and David R. Holian The Dynamics of Committee Outliers: Evidence from the House of Representatives, Journal of Legislative Studies 3 (Summer): Cox, Gary W. and Matthew D. McCubbins Legislative Leviathan: Party Government in the House. Berkeley and Los Angeles: University of California Press. Davidson, Roger H Subcommittee Government: New Channels for Policy Making. In Thomas E. Mann and Norman Ornstein, eds., The New Congress. Washington: American Enterprise Institute. Davidson, Roger H. and Walter J. Oleszek Congress and Its Members. Washington: CQ Press. Sixth edition. Deering, Christopher J. and Steven S. Smith Committees in Congress. Washington: CQ Press. Third edition. Fenno, Richard F., Jr The Power of the Purse: Appropriations Politics in Congress. 28

30 Boston: Little, Brown. Fenno, Richard F., Jr Congressmen in Committees. Boston: Little, Brown. Francis, Wayne L The Legislative Committee Game: A Comparative Analysis of Fifty States. Columbus: The Ohio State University Press. Gilligan, Thomas W. and Keith Krehbiel Organization of Informative Committees by a Rational Legislature. American Journal of Political Science 34 (May): Groseclose, Tim Testing Committee Composition Hypotheses for the U. S. Congress. The Journal of Politics 56 (May): Hall, Richard L. and Bernard Grofman The Committee Assignment Process and the Conditional Nature of Committee Bias. American Political Science Review 84 (December): Hearings on Committee Organization in the House Before the Select Committee on Committees United States House of Representatives, 93 rd Congress, 1 st session, III. Herron, Michael C Artificial Extremism in Interest Group Ratings and the Preferences versus Party Debate. Legislative Studies Quarterly 24 (November): Kiewiet, Roderick and Matthew D. McCubbins The Logic of Delegation: Congressional Parties and the Appropriations Process. Chicago: University of Chicago Press. Krehbiel, Keith Are Congressional Committees Composed of Preference Outliers? American Political Science Review 84 (March): Krehbiel, Keith Information and Legislative Organization. Ann Arbor: University of Michigan Press. Krehbiel, Keith Deference, Extremism, and Interest Group Ratings. Legislative Studies 29

31 Quarterly 19 (February): Liebowitz, Arleen and Robert Tollison A Theory of Legislative Organization: Making the Most of Your Majority. Quarterly Journal of Economics 94 (March): Londregan, John and James M. Snyder Comparing Committee and Floor Preferences. Legislative Studies Quarterly 19 (May): Maltzman, Forrest Meeting Competing Demands: Committee Performance in the Postreform House. American Journal of Political Science 39 (August): Manley, John F The Politics of Finance: The House Committee on Ways and Means. Boston: Little, Brown. Marshall, Bryan W., Brandon C. Prins, and David W. Rohde Theories of Legislative Organization: An Empirical Study of Committee Outliers in the Senate. Working Paper, Political Institutions and Public Choic Program, Michigan State University, East Lansing, MI. Mayhew, David Congress: The Electoral Connection. New Haven: Yale University Press. McConachie, Lauros G Congressional Committees: A Study of the Origins and Development of Our National and Local Legislative Methods. New York: Thomas Y. Crowell. Overby, L. Marvin and Thomas A. Kazee Outlying Committees in the Statehouse: An Examination of the Prevalence of Committee Outliers in State Legislatures. The Journal of Politics 62 (August): forthcoming. Polsby s testimony Poole, Keith T. and Howard Rosenthal Patterns of Congressional Voting. American 30

32 Journal of Political Science 35 (February): Price, David E Policy Making in Senate Committees: The Impact of Environmental Factors. American Political Science Review 72 (June): Rohde, David W Parties and Leaders in the Postreform House. Chicago: University of Chicago Press. Shepsle, Kenneth A The Giant Jigsaw Puzzle. Chicago: University of Chicago Press. Shepsle, Kenneth A. and Barry R. Weingast Positive Theories of Congressional Institutions. In Kenneth A. Shepsle and Barry R. Weingast, eds., Positive Theories of Congressional Institutions. Ann Arbor: University of Michigan Press. Snyder, James M Artificial Extremism in Interest Group Ratings. Legislative Studies Quarterly 17 (August): Squire, Peverill Legislative Professionalization and Membership Diversity in State Legislatures. Legislative Studies Quarterly 17 (February): Weingast, Barry R. and William Marshall The Industrial Organization of Congress. Journal of Political Economy 96 (February): Wilson, Woodrow Congressional Government. Boston: Houghton Mifflin. 31