What Happens When Insurers Make Insurance Laws? State Legislative Agendas and the Occupational Makeup of Government

Abstract

Do the occupational backgrounds of politicians affect the government’s agenda? Businesses have long thought so. The first occupational data on state legislators were collected by the Insurance Information Institute, an interest group representing major insurance companies. In this paper, we test one potential motive for these kinds of efforts: the idea that the occupational makeup of governments affects the agendas they pursue, an argument that has been largely neglected in research on politicians’ occupational backgrounds. We focus here on the insurance industry. Using original data, we find that state legislatures with more former insurers consider fewer bills regulating insurance (negative agenda control), that former insurers play a disproportionate role in drafting the insurance bills that are introduced (positive agenda control), and that the bills former insurers introduce tend to be more favorable to the industry than those that their colleagues introduce (positive agenda control). The occupational makeup of legislatures may indeed affect their agendas, as industry groups have long suspected.

Keywords

descriptive representation insurance social class occupation agendas state politics

Does it matter what professional backgrounds politicians come from? Should scholars and activists pay attention to the occupational makeup of political institutions? Industry organizations have long thought so. The first people to collect recurring national data on the aggregate occupational makeups of state legislatures in the United States were not scholars of American politics but researchers at the Insurance Information Institute, an interest group representing dozens of major national insurance companies. In the 1960s and 1970s, the Institute began collecting nationwide data on the occupations that state lawmakers came from, including the number who were actively employed in the insurance industry—presumably in the hopes that those legislators would be especially favorable to policies that benefitted insurers.

Was their hunch right? Do politicians tend to favor the industries they worked in? As it stands, scholars of U.S. politics have only limited evidence. Most research on the personal roots of elite decision making (Burden 2007) and the numerical or descriptive representation of social groups (Pitkin 1967) has focused on characteristics such as race and gender (e.g., Berkman and O’Connor 1993; Canon 1999; Griffin and Newman 2008; Swers 2002; Thomas 1991; Whitby 1997). Only a handful of studies have explored whether politicians from different occupations behave differently in office (Carnes 2012, 2013, 2016; Eulau and Sprague 1964; Miller 1995; Witko and Friedman 2008). And although their findings generally suggest that occupations matter, these studies still leave many stones unturned.

Perhaps one of the most pressing weaknesses in the emerging literature on politicians’ occupational backgrounds is that most studies have focused on the associations between a legislator’s former occupation and how he or she casts roll-call votes. This is a logical place for researchers to begin; data on roll-call voting are easy to obtain and analyze. But just as the literatures on the gender and racial backgrounds of politicians moved from early roll-call analyses to research on the consequential but harder-to-study choices that legislators make before the votes are cast, it is probably time for the literature on the occupational backgrounds of politicians to move beyond the low-hanging fruit.

In this paper, we ask whether politicians’ occupational backgrounds are associated with the important work that goes on during the agenda-setting stages of the legislative process (Hall 1996; Kingdon [1984] 2011; Schattschneider 1960). Does the presence or absence of politicians from a given occupational background affect the kinds of bills that are introduced in a legislature? As a test of the hypothesis that lawmakers’ occupations can inform the legislative agenda, we focus on the trillion-dollar industry that pioneered quantitative research of the occupational makeup of American legislatures: the insurance industry. From an electoral standpoint, insurance policy can be an unrewarding issue. Insurance-related legislation is often highly technical, the vast majority of insurance legislation receives little or no media attention, and routine insurance bills rarely excite the passions of the general public.¹ The lawmakers who work on insurance issues are, therefore, likely to be those who already have expertise and interest in the field—including, we suspect, legislators who have worked in the insurance industry themselves.

Drawing on data from three recent state legislative sessions (2007–2008, 2009–2010, and 2011–2012), we study the 30 states for which we could obtain both information about legislators’ occupational backgrounds (either their main occupation before holding office, or for those who were still employed, their main occupation outside of holding office) and data on the bills introduced in each state that affected the insurance industry. Using a mix of human coding and automated text analysis, we classified the insurance bills introduced in these states based on whether they benefitted the insurance industry.

With these data, we then explore whether politicians from the insurance industry affect the legislative agenda on insurance regulation, either by promoting favored legislation (positive agenda control) or by blocking unwanted regulations (negative agenda control). Our analyses find evidence of both. States with more insurance industry employees in their legislatures consider fewer bills regulating the industry overall (Study 1). But the bills that are introduced are disproportionately influenced by former insurers (Study 2), and the bills insurers introduce, while not overwhelmingly favorable to the industry, tend to be more favorable than their colleagues’ insurance bills (Study 3).

These findings have important implications for the study of descriptive and substantive representation and the personal roots of elite decision making, and especially the emerging research on the occupational makeup of governments in the United States and elsewhere. This study adds new empirical evidence to the growing finding that politicians’ occupational backgrounds matter. It studies an occupational group that has never been the subject of past research on politicians’ occupational backgrounds; just as scholars have shown that politicians from law, business, and working-class jobs differ from other leaders, this study shows that politicians from insurance seem to stand out as well. Most importantly, this paper is one of the first studies to focus exclusively on how occupational backgrounds matter in the agenda-setting process, not just during the roll call votes that have more often been the subject of academic research. And this study goes further than past research on occupations and agendas by taking seriously the distinction between negative and positive agenda control, a point that has so far been overlooked in the emerging research on the occupational backgrounds of politicians. This study supports the emerging consensus that it does, in fact, matter what professional backgrounds politicians come from—and suggests that scholars interested in this topic would do well to focus more carefully on how the occupational makeup of political institutions can affect which problems and policies ultimately make it onto the government’s agenda.

Agenda Setting and Occupations

The legislative agenda—the set of problems lawmakers hope to act on and the set of policies they are considering to solve them—is one of the most important sites of power in the policymaking process. If lawmakers are unwilling to entertain a given policy, it cannot be considered or debated, let alone enacted. If they are uninterested in a problem, government interventions to solve it stand little chance of getting off the ground.

Scholars of legislative politics have long recognized that the government’s agenda depends at least in part on the personal views and backgrounds of legislators. Of course, crises, unexpected events, and sudden shifts in public opinion sometimes force problems or policies into the spotlight. Lawmakers sometimes sponsor bills, work on committees, and participate in floor debates in the hope of building reputations that constituents, donors, interest groups, and party leaders will view favorably (Hansen and Treul 2015; Koger 2003; Kollman 1998; Weingast and Marshall 1988). However, lawmakers also routinely champion causes that they have some special expertise in, or some personal connection to (Burden 2007; Fenno 1973; Hall 1996; Kingdon [1984] 2011; Krehbiel 1991; Mayhew 2000; Schiller 1995; Wawro 2000). That is, lawmakers often fight for proposals that they know more about or care more about.

The proposals lawmakers know and care about are in turn often correlated with their personal backgrounds and life experiences (Burden 2007). For example, black lawmakers work harder than their white colleagues to promote legislation on issues that affect the black community (Bratton and Haynie 1999; Grose 2011) while women in legislatures work harder to promote legislation on women’s issues (Little, Dunn, and Deen 2001; Swers 2002). The numerical or descriptive representation of many social groups seems to affect whether their concerns are mirrored in the legislative agenda.

Is the same true for occupational groups? Do legislators from different lines of work champion different kinds of bills? Occupational backgrounds differ from other demographic and social characteristics in important ways; race and gender are not the same as occupation (most notably, because occupations are far more fluid and self-selected). However, there are still theoretical reasons to expect politicians from different occupational backgrounds to know and care about different subjects (albeit for reasons that have less to do with experiences rooted in immutable characteristics). People from different occupations often tend to have different views, especially on issues that directly pertain to their industries. This may be because they have different material or political incentives; a simple capture model would predict that legislators from a given industry would have personal financial incentives (e.g., the revolving door) or political incentives (e.g., campaign contributions from lobbyists) to promote new laws that favor the industry, as scholars of bureaucracy recognize (see Carpenter 2014).

Alternatively, lawmakers’ careers might simply socialize them to prioritize different problems and to think differently about proposed solutions. Cultural capture models would predict that legislators from a given industry would work to promote it, not because of personal material interests but because their experiences in the industry shaped their perceptions about what constitutes good public policy (Jansa and Gray 2017; Kwak 2014). Or lawmakers might even be motivated to act on behalf of their industries due to a more conscious feeling of shared identity with their occupational group.² A retired public school teacher serving in the legislature might fight to increase teacher salaries, not because she expects to benefit personally, but because her time working as a teacher influenced her personal opinion on how much teachers should be paid or left her with a feeling of solidarity with today’s educators.

Whatever the exact mechanism, lawmakers from different occupations may well bring different priorities to the agenda-setting process. Do they?

Advancing Research on Lawmakers’ Occupational Backgrounds

To date, there has been almost no research on the links between the occupational backgrounds of politicians and the agendas that political institutions pursue. In general, research on the descriptive representation of occupational groups has been somewhat sporadic. In the 1960s, there was a burst of scholarship on the careers and social circles politicians were drawn from (e.g., Domhoff 1967; Matthews 1964), but this line of research slowed in the 1970s (for a review, see Putnam 1976) and was more or less dormant until the early 2000s (but see Miller 1995). In the last decade, scholars have taken a renewed interest in this topic (e.g., Bonica 2017; Campbell and Cowley 2014; Carnes 2013; Carnes and Lupu 2015; Sojourner 2013; Witko and Friedman 2008). However, this new wave of research on politicians’ occupational backgrounds has still left many questions unanswered. Scholars have only examined a handful of occupations. For instance, we know something about politicians from law (Bonica 2017; Miller 1995), business (Witko and Friedman 2008), farming (Bellemare and Carnes 2015), and working-class jobs (Carnes and Hansen 2016), but not other lines of work.

Perhaps more problematic, the recent work on the effects of politicians’ occupational backgrounds (or related measures like their wealth or outside income) has focused almost exclusively on just one form of legislative conduct, namely, how lawmakers cast roll call votes (Carnes 2012, 2016; Eulau and Sprague 1964; Griffin and Anewalt-Remsburg 2013; Kraus and Callaghan 2014; Witko and Friedman 2008). This is an understandable choice. Roll-call voting is far easier to measure than agenda-setting activities like introducing bills (Hall 1996). Moreover, many studies have found important links between occupation and roll calls. Just as ordinary citizens from different occupations often have different views about economic issues, lawmakers from different occupational backgrounds tend to vote differently on economic legislation. However, contemporary research has stopped short of asking whether the occupational differences in how legislators vote are also apparent when they do the important work of crafting bills, advocating problems, and setting the legislative agenda.

The rare studies that have broached the subject have had significant (and admitted) methodological limitations. One approach has been to focus on the issues lawmakers work on, but not the stances they take. For instance, Carnes (2013, chap. 3) used data on the bills members of Congress introduce and the numbers of cosponsors they recruit to show that legislators from working-class occupations work harder to promote domestic economic policies. However, as the study notes, the research design could only identify the issues each bill addressed (e.g., economic policy vs. agricultural policy), not the actual policy implications of each proposal (e.g., pro-worker vs. pro-business proposals).

A second approach has been to ask whether legislators are disproportionately likely to serve on committees that regulate the industries they work in (or worked in in the past). This research has found clear evidence that legislators seek out committee assignments related to their current or former industries (e.g., Battista 2012; Buchanan 1962; Hamm, Hedlund, and Post 2011; Renzulli and Center for Public Integrity 2002). But as these studies often note, data on committee appointments are not the same as data on the actual policies or proposals legislators pursue. In principle, at least, a lawmaker who worked in health care might use a seat on a health-related committee to increase taxes and regulation on health care providers.

Whereas most prior research on the occupational backgrounds of politicians has focused on committee assignments or numbers of bills introduced, in this paper, we analyze the links between legislators’ occupations and both the volume and direction of new legislation on relevant issues. That is, we go beyond past research, which has asked whether legislators from a given occupational background are more likely to introduce related bills or join related committees. We ask whether legislators from a given occupational background are more likely to give their former or current industry not just more legislation but the kind of legislation that it wants.

This approach has the advantage of allowing us to study both negative and positive agenda control. When a group or industry wants to exert positive agenda control—getting preferred policies or problems onto the government’s agenda—existing measures of legislative effort (number of bills, number of cosponsors, and so on) should capture that phenomenon, albeit somewhat imprecisely. But for many groups and industries, negative agenda control—preventing undesirable ideas or proposals from ever getting serious consideration in government (Bachrach and Baratz 1962; Baumgartner et al. 2009; Schattschneider 1960)—is just as important.

If a social group or industry wants a mix of both positive and negative agenda control, studying legislative effort alone may lead us astray. Suppose, for instance, that farmers-turned-legislators fight harder than other members for new farm subsidies but are also less likely than other members to introduce bills regulating food safety in agricultural production. In this case, a simple measure of total legislative effort might suggest that farmer-legislators introduce about as many agricultural bills as other members on average—and lead researchers to wrongly conclude that farmer-lawmakers are just like other members. By studying both the aggregate volume of legislation introduced on a given issue (which can capture instances when negative agenda control outweighs positive agenda control, or vice versa) and the direction or substance of individual bills (which can be either beneficial or harmful to a given industry), our research captures a wider range of agenda-setting behaviors.

Insurance and State Legislative Agendas

To test the idea that lawmakers’ occupational backgrounds influence their agenda-setting activities, we focused on state legislators who had previously worked in the insurance industry. Why insurance? For one, the industry is large and important. The United States has the largest private insurance industry in the world, and its annual revenues usually exceed $1.2 trillion. Over two million Americans are employed in insurance, and insurance makes up about 40% of the financial sector (Millard 2015).

Second, the insurance industry has long sought to influence public policy, and a major interest group representing the industry, the Insurance Information Institute, collected data on the occupational backgrounds of U.S. politicians in the hopes of doing so. In that sense, it is a good starting point for research on occupational backgrounds and agenda setting and, in fact, a crucial case for our argument. If any occupational background influences legislative agenda setting, it should be a background in insurance.

Third, and related, we suspect that if legislators from insurance backgrounds somehow influence the agenda-setting stage, legislators from other large, well-organized industries might have similar effects on agenda-setting that could be investigated in future research. The insurance industry is deeply influenced by public policy and highly attentive to who holds office. Of course, many politicians hail from occupational backgrounds in industries that are less well-organized and less focused on government (e.g., legislators working in the nonprofit or service sectors). But politicians from any large, well-organized industry that regularly monitors and courts elected officials—real estate, agriculture, finance, health care, and so on—may well exhibit similar behaviors.

Do lawmakers who worked in insurance differ in how they attempt to shape the legislative agenda? Do they try to make the insurance industry’s agenda the government’s agenda? In this paper, we focus on state legislatures. States are ideal for our purposes for several reasons. State agendas are 50 times more numerous than the federal agenda and far more varied. The occupational makeups of state legislatures also vary far more than the makeup of federal offices (Carnes 2013). Unlike county and local governments, states have significant agenda-setting powers, and their agendas can be studied far more readily.

And—perhaps most importantly—state governments make many of the most important policy decisions pertaining to the insurance industry. States license insurance companies and individual insurance producers (also known as insurance agents or brokers). They also regulate insurance markets, insurance company finances, and insurance products (often setting caps on premiums and defining crucial concepts like what constitutes “reasonable and fair” information or which policies contain gaps that might be misunderstood by consumers). States have remained the primary drivers of insurance regulation even after the significant changes made to insurance markets under the Affordable Care Act.³ Often, individuals and organizations must purchase insurance policies from in-state providers who conform their policies to meet state requirements. If a legislator’s occupation matters for any issue, it should be insurance—and if an insurance background matters in any institutional context, it should matter in state legislatures.⁴

Unfortunately, the Insurance Information Institute no longer publicly shares its data on the aggregate occupational makeup of state legislatures (and has never shared its data on the occupations of individual lawmakers). To determine whether legislators from insurance backgrounds were more likely to promote insurance’s agenda in the legislature, we collected data describing the occupational makeup of each state’s legislature over the span of two decades from state legislative websites, online historical databases, published legislative manuals, and state directories.⁵ These occupational data captured the main occupations that each legislator held outside of office, both while they were serving (for state lawmakers who hold outside jobs) and before (for all lawmakers).⁶ To ensure we were using reliable occupational data, we limited our collection effort to the states that recorded lawmakers’ occupations and published that information online or in print. Although we omitted 20 states where officials never collected occupational data or do not maintain historical records, we were more certain about the validity of these data than the data we could have collected from online web searches.

Of course, an analysis of agenda-setting on insurance must be sensitive to both negative and positive forms of agenda-setting. Insurers benefit from some regulations (like the requirement that drivers purchase auto insurance) and shoulder burdens from others (like fines levied to brokers who fail to meet continuing education requirements). On balance, insurers probably favor fewer new regulations overall. But insurers also regularly advocate reforms, changes to licensing requirements, and many other interventions. Against this backdrop, a simple measure of total legislative effort would miss essential features of the way the legislative agenda helps and harms insurance companies. As such, we sought to measure both legislative effort and legislative content.

Using our occupational data, we conducted three related studies to measure (1) the aggregate-level relationship between the number of lawmakers from insurance backgrounds and the amount of legislative effort devoted to insurance issues in the state legislature, (2) whether individual lawmakers from insurance backgrounds were disproportionately likely to introduce or influence the drafting of insurance legislation, and (3) whether the insurance bills introduced by lawmakers from insurance backgrounds were more likely to be favorable to the industry.

On balance, we expected to find a negative relationship between the presence of lawmakers from insurance backgrounds and the total volume of legislative effort devoted to insurance. Government regulation is often a burden to the industries being regulated, and we expected to find evidence that lawmakers from insurance backgrounds would exercise some negative agenda control. The more lawmakers from insurance who became state legislators, we reasoned, the less the state legislative agenda would focus on regulating the insurance industry.

Beneath this larger pattern, however, we expected to find evidence of positive agenda control as well. Among the insurance bills that are introduced, we expected former insurers to be heavily involved, and we expected them to be disproportionately favorable to the industry. When lawmakers from insurance backgrounds introduced insurance bills, we expected those bills to be more favorable to insurers (relative to the insurance bills introduced by other members). Large, complex industries have incentives to block some regulations and promote others. If lawmakers’ occupational backgrounds influence their work at the agenda-setting stage, we should expect legislators from the insurance industry to generally work to keep regulation off the agenda—but also to champion the proposals that insurers want.

Study 1: Do State Legislatures with More Insurers Consider Fewer Insurance Bills?

We first examined the aggregate-level relationship between the share of insurers in a state’s legislature and the volume of new legislation that focused on the insurance industry. Following Gray et al. (2005), we used Lexis-Nexis subject term searches to count the number of bills and bill amendments⁷ that addressed insurance-related topics in each state legislature in 2007, 2009, and 2011.⁸ We chose these years based on the availability of concurrent occupation and bill count data.

Using our individual-level legislator occupation data, we first created a simple measure of the proportion of lawmakers in each state who had previously or currently worked in the insurance industry. Because our occupational data captured the main or most important occupations a legislator worked, if a legislator reported working in insurance and in other occupations, we still counted him or her as having a professional background in insurance (e.g., a legislator who listed “insurance broker/real estate agent”—the most common compound occupation—would still be counted as having worked in insurance). Of course, this approach could potentially code legislators as insurance professionals even when they only spent a short time working in the industry and spent most of their career elsewhere. Unfortunately, most states do not report how long legislators spent working in each industry. If career professionals are more likely to advance their industry’s agenda, then our coding scheme should work against any positive findings by including individuals with weaker personal attachments to the industry in the data.

We also collected data on several control variables. Because the number of bills introduced can vary drastically from state to state, we collected data on the total number of bills introduced in each state from The Book of the States. Importantly, when we counted the number of bills introduced on each issue using Lexis-Nexis, our search results counted both bills and amendments. The Book of the States, on the contrary, counted each numbered bill only once. Even so, we considered this a reasonable way to control for state-to-state differences in the total size of the legislative agenda.⁹

State lawmakers might introduce more legislation related to a certain industry when the industry makes up a greater share of the state’s economy. For example, Iowa is likely to consider more agriculture legislation each year than Connecticut. As such, we controlled for the size of the insurance industry in each state using American Community Survey data on the total proportion of citizens employed by insurers.¹⁰ We also recorded the proportion of the chambers in each legislature that Democrats controlled.¹¹ And to account for differences in how actively legislators were lobbied about insurance issues, we controlled for the lobbying capacity of the insurance industry in each state using data on the number of interest groups lobbying on insurance issues in 2007 (Lowery et al. 2012).

Figure 1 plots the basic relationship between the number of insurance-related bills and amendments introduced in each state (on the vertical axis) and the share of state lawmakers with professional experience in the insurance industry (on the horizontal axis). Consistent with our expectations, in states with more insurers in their legislatures, insurance made up a smaller share of the legislative agenda.

Figure 1.

Do state legislatures with more insurers consider less insurance legislation?

Of course, the occupational makeup of government could be associated with a wide range of other factors that might drive the legislative agenda. Maybe states with lots of insurers in their legislatures are also states where insurance is a major industry, where Republicans hold more seats, where insurers lobby legislators more vigorously, or where legislatures are more professionalized. To check for these kinds of confounding variables, we estimated three negative binomial regression models reported in Table 1, one without controls, one with controls for the year, and the total number of bills introduced in each state legislature, and one with all the controls described previously.

Table 1.

Negative Binomial Regression Models Relating the Volume of Insurance Legislation to the Share of State Lawmakers from Insurance Backgrounds.

	(1)	(2)	(3)
Percentage of lawmakers from insurance	−0.11*	−0.03	−0.14*
Percentage of lawmakers from insurance	(0.05)	(0.06)	(0.07)
Percentage of population employed in insurance			−0.14*
Percentage of population employed in insurance			(0.06)
Democratic control of state legislature (ind.)			−0.48^†
Democratic control of state legislature (ind.)			(0.28)
Number of insurance lobbying groups			0.00
Number of insurance lobbying groups			(0.00)
Legislative professionalism			−0.04
Legislative professionalism			(0.06)
Year fixed effects	No	Yes	Yes
Total number of bills (offset)	No	Yes	Yes
Intercept	6.50*	3.25*	4.40*
	(0.14)	(0.20)	(0.59)
N	89	89	83
BIC	1,281.48	1,243.07	1,163.97

Source. Authors’ data collection.

Note. Standard errors in parentheses in Model 1. Robust standard errors clustered by state in parentheses in Models 2 and 3. Missing observations in Model 3 are due to missing data for party control in Nebraska and legislative professionalism. BIC = Bayesian Information Criterion.

†

p < .10. * p < .05, two tailed.

Like Figure 1, the models related the number of insurance bills and amendments introduced in each state legislature to the percentage of lawmakers who had worked in insurance. The first model simply regressed the size of the insurance agenda in each state on the share of legislators from insurance backgrounds. As in Figure 1, the relationship was negative and statistically significant; compared to a state with no insurers in its legislature, a state legislature with 5% insurers considered roughly 270 fewer insurance bills and amendments on average.¹²

In the second model, we added the total number of bills introduced in each legislature (as an offset variable) to account for the fact that some states consider more bills in general. We also included year fixed effects, and we clustered standard errors to account for repeated observations of states over time. In the third model, we also added controls for the partisan makeup of the legislature (whether the legislature was controlled by Democrats), lobbying on behalf of the insurance industry (the number of registered groups lobbying on insurance matters in the state), potential pressures from citizens employed in the insurance industry (the percentage of the states’ citizens who were employed in insurance), and legislative professionalism. In model 2, the estimated association between occupations and agendas was still negative and substantively large—according to the model, compared with a state with no insurers in its legislature, a state legislature with 5% insurers considered roughly 118 fewer insurance bills and amendments on average—although the association dropped below statistical significance. In model 3, the association was statistically significant and substantively comparable with model 1, suggesting that states with 5% insurers considered roughly 472 fewer insurance bills and amendments on average relative to states with no insurer-legislators. (The average number of insurance bills considered by a state in this sample was 516.)¹³

Figure 2 plots predictions from the models in Table 1 graphically. Viewed this way, it is easy to see not only the differences across each of the three models’ results but also the larger pattern that is common across the three. In times and places where people from the insurance industry hold more seats in state legislatures, legislatures introduce fewer bills related to insurance. Even when we take into account the size of legislative agenda in each state (models 2 and 3) and the role of insurance professionals in the general public and the organized pressure group system (model 3), the presence of insurance professionals in the statehouse is associated with significant reductions in the attention legislators pay to the insurance industry.

Figure 2.

When insurers hold more seats, legislatures consider fewer insurance bills.

To check that the results of this analysis generalized beyond the 30 states for which we have occupational data, we also re-ran model 3 using data from the National Conference of State Legislatures, which collected aggregate-level data on the percentage of lawmakers from the insurance industry in each state’s legislature in 2007 (see Table A2 in the online appendix); the association between insurers and insurance bills in this model was statistically indistinguishable from the association in model 3 in Table 1.¹⁴ To ensure that our findings were not simply a fluke of the way we modeled count data, we re-estimated model 3 in Table 1 using ordinary least squares (OLS) and Poisson regression, and to check that the cases in the upper left corner of Figure 1 were not driving our results, we re-estimated model 3 omitting the observations for California 2007, New York 2007, and New York 2009 (Table A3 in the online appendix). All three models again found negative associations, and only the OLS model fell short of conventional thresholds for statistical significance (p < .186). To account for the fact that the volume of new insurance legislation might depend in part on how much insurance legislation was passed in previous legislative sessions, we re-estimated models 2 and 3 in Table 1 adding controls for the numbers of insurance bills introduced in previous sessions; our main findings were unchanged (see Table A4 in the online appendix). Finally, because our data come from legislative sessions that overlap with the passage of the Affordable Care Act at the federal level in 2010, we estimated bivariate negative binomial regression models separately for each year in Table A5 in the online appendix to ensure that our results were consistent across each of the individual legislative terms in our dataset. These models found negative associations between the independent and dependent variables in each year, although in two of the years, the point estimates were not statistically significant, perhaps because of the small number of observations (30) for each year. Overall, all of our follow-up models found negative associations between insurers and the volume of new insurance laws in state legislatures.

With a small sample and a blunt measure of the state legislative agenda, our data suggested that legislatures with more former insurers tend to spend less time considering insurance regulations. Importantly, these results do not necessarily mean that legislatures with more former insurers introduce fewer insurance bills; our measure of the agenda includes both bills and amendments. As such, these findings suggest that when insurers hold more seats, legislatures spend less time overall debating and regulating insurance—consistent with the idea that former insurers exercise negative agenda control.

Study 2: When Insurance Bills Are Proposed, Do Insurers Play an Outsized Role?

Insurers may not want much government regulation overall. But when it comes to the bills that insurers do want, do lawmakers from the insurance industry exert positive agenda control? One way to find out is to study who influences the insurance bills that are introduced. If insurers use their positions in legislatures to shape the agenda, we would expect not only that they would discourage many bills (contributing to an overall decline in the number of insurance laws introduced) but also that they would play an outsized role in the insurance bills that are introduced.

To find out whether this is the case, we turned to more fine-grained data from Open States, an independent, open-source data-collection project previously run by the Sunlight Foundation.¹⁵ Open States aggregates data on state legislative bills, legislators, and actions directly from state legislative websites and makes the data available for bulk downloads using their application program interface (API). Using Open States’ data, we searched for all legislation concerning the insurance industry introduced in the 30 states for which we have occupational data, this time focusing on the 2011–2012 legislative term (for which Open States has the most complete data on the bills introduced in each state).¹⁶

Using the search term insur in API’s search parameter, we first identified the 3,706 bills related to insurance regulation introduced in our 30 states during the 2011–2012 legislative term.¹⁷ We then downloaded information about each bill, including its state and chamber of origin, bill number, title, full-text URL, and its sponsors and cosponsors. We then matched each bill’s sponsors and cosponsors to our list of insurance professionals, and we coded each bill for whether insurance professionals were involved in its introduction. For bills introduced by committees (not individual legislators), we coded whether an insurance professional was a member or chair of the committee.

Table 2 breaks down the different ways that insurance professional might have been involved in introducing insurance legislation.¹⁸ Consistent with our expectations, former insurers seemed to play a disproportionate role in the insurance legislation that made it onto state legislatures’ agendas. Of the more than 3,700 insurance bills that were introduced in the 30 states, 234 (6.3%) were sponsored by an insurance professional, 153 were cosponsored by an insurer (4.1%), and another 99 (2.7%) were sponsored by committees that were either chaired by insurers or had insurers among their members. Whereas insurers made up between 2% and 3% of the typical legislature (and other subgroups of legislators that might have disproportionate influence over insurance legislation, like Republicans, men, white legislators, and legislators with more than 10 years of experience), insurer-influenced bills made up more than 13% of the insurance legislation introduced during this time period. That is, insurers were more than five times as likely to be involved in crafting insurance bills as we would expect based on their numerical representation. Even looking only at bills’ primary sponsors, insurers were more than twice as likely to write an insurance bill as we would expect by chance alone.

Table 2.

Are Insurer-Legislators More Involved in Insurance Bills?

Insurance professional was…	No. of bills	% of bills
Primary sponsor	234	6.31
Cosponsor	153	4.13
Chair of sponsoring committee	54	1.46
Member of sponsoring committee	45	1.21
Total	486	13.11
Average percent of insurers in legislature		2.56
Average percent of insurers among Republican legislators		3.17
Average percent of insurers among white legislators		2.84
Average percent of insurers among male legislators		3.07
Average percent of insurers among legislators with 10+ years’ experience		3.02

Source. Authors’ data collection.

Note. Rows are treated as mutually exclusive in descending order. That is, if insurers were among the primary sponsors and cosponsors on a bill, it would be counted among the bills that had insurers as their primary sponsors.

Of course, the vast majority of bills that affect the insurance industry—87%—are introduced by lawmakers who do not have backgrounds in the insurance industry. In that sense, concerns about industry capture of insurance regulation are probably overstated. But the occupational background of a legislator still seems to matter in this instance. Insurers may not have absolute control over state-level insurance legislation, but they have an outsized role in the insurance bills that are introduced in state legislatures, consistent with the idea that they exert some positive control over the state insurance agenda.

Study 3: Do Legislators from Insurance Backgrounds Sponsor Industry-Friendly Bills?

Simply knowing that insurance professionals are disproportionately involved in insurance regulation does not necessarily mean that they are steering the regulatory agenda toward outcomes that the industry favors. What kinds of insurance bills do former insurers introduce? To find out, we analyzed the content of the insurance bills identified in Study 2 using a combination of human coding and automated text analysis. We first collected URLs for the full text of all 3,706 of the insurance bills introduced in the 2011–2012 legislative in the 30 states for which we have individual-level data on whether lawmakers worked in insurance. We then used web scraping software to download the full text of each bill. Broken links and incompatible file formats prevented us from collecting every piece of legislation (and cost us two states entirely, Massachusetts and Oklahoma), but we were ultimately able to gather the text of 3,169 bills (86% of the sample) from 28 states.

If legislators’ occupations influence how they behave at the agenda-setting stage, we would expect insurance professionals to sponsor more legislation that moves insurance policy in an industry-friendly direction and less legislation that moves policy away from industry preferences. Of course, insurance regulation is complex and spans many types of financial products. A bill may contain many different provisions, some favored by the industry and others opposed. In general, however, most insurance bills tend to change government involvement in insurance markets in ways that are either favorable or unfavorable to insurers.

Due to the inherent complexity of legislative text, we did not expect an unsupervised text classification program—one that codes texts in a completely automated fashion—to effectively distinguish between pro-industry and anti-industry bills. Instead, we used a supervised classifier, a program that relies in part on human coding. Of the 3,169 bills, we randomly sampled 500 to code ourselves.¹⁹ Our overarching principle for determining whether a bill was pro- or anti-industry was how the bill affected insurance companies’ bottom line. Pro-industry bills included those that used government funds to help individuals purchase insurance, bills that increased government sanctions for consumer insurance fraud, and bills that decreased insurance companies’ liability limits. Anti-industry bills included those that required insurers to offer policies covering certain types of medical procedures or services, bills that increased insurers’ burden for documenting compliance with regulations, or bills prohibiting business practices deemed unfair by consumer protection advocates. When possible, we searched for media accounts that mentioned the support or opposition of industry groups. When bills were not obviously pro-industry or anti-industry, or did not seem to us to be pertinent to private insurance regulation (e.g., state unemployment insurance bills), we simply coded them as Unknown.

After coding these 500 bills, we then categorized them using multinomial least absolute shrinkage and selection operator (LASSO) regression (Tibshirani 1996) to predict whether each bill was pro-industry, anti-industry, or unknown.²⁰ That is, we ran a text-classification model that used the language in 400 of our coded bills to identify phrases that distinguished between pro- and anti-industry bills (and unknown bills), then tested the model by using the text of the remaining 100 coded bills to generate text-based predictions of whether the bills were pro- or anti-industry. The model accurately predicted the direction of the bill 73% of the time,²¹ a fairly high accuracy for complicated texts like insurance regulation bills.²² (Moreover, misclassification will only reduce the efficiency of our analyses, thereby reducing the chances that we will find that insurance professionals sponsor more insurance-friendly portfolios of bills.) We then used our trained model to code the remaining 2,669 bills in our sample.

Using these data on the direction of each bill, we then asked whether insurers were more likely to sponsor or cosponsor bills that were favorable to the insurance industry, and whether insurers were less likely to sponsor bills that were unfavorable to the industry. The results are presented in Table 3. Several patterns immediately stand out. First, consistent with our idea that new insurance legislation is more often unfavorable to industry, a majority of the coded insurance bills in our sample—57%—were unfavorable to insurance companies. Legislators attempting to protect the insurance industry’s interests have good cause to try to reduce the overall volume of new insurance legislation (as we saw in Study 1).

Table 3.

Do Insurers Introduce More Pro-Industry Bills?

	Bills backed by insurers	Bills backed by other legislators	Total
Pro-industry	191 (48.8%)	1,041 (41.7%)	1,232 (42.7%)
Anti-industry	200 (51.2%)	1,454 (58.3%)	1,654 (57.3%)
Total	391	2,495	2,886

Source. Author’s data collection.

Note. Cells report counts (with column percentages in parentheses). In a t test, the difference between insurers’ bills and other bills was statistically significant at p < .008.

Second, bills that insurers sponsored were almost equally likely to be favorable or unfavorable to the industry. Setting aside bills coded as Unknown (which made up 8.8% of former insurers bills and 9.7% of non-insurers’ bills), roughly 49% of insurers’ bills were favorable, while 51% were unfavorable to the industry. This finding lies in contrast with our expectation that bills sponsored by former insurers would be more likely than not to be friendly to industry.

However, insurers were significantly more likely than their colleagues to introduce bills that were favorable to the industry. The bills for which insurers were sponsors, cosponsors, or chairs/members of sponsoring committees were about seven percentage points more likely to be pro-industry than the insurance bills introduced by non-insurers (48.8% vs. 41.7%; diff. = 7.1%; p < .008). This gap, moreover, may understate the true differences between insurers and other legislators slightly. In our 500 human-coded bills, for instance, insurer-backed bills were nine percentage points more likely to be favorable to industry (54.2% vs. 45.2%; diff. = 9.0%; p < .20).²³

There are many factors besides a legislator’s occupational background that might influence the kinds of insurance bills he or she introduces. Could the previous results be driven by some other characteristic common to legislators from insurance backgrounds? To find out, we estimated two logistic regression models. The first simply regressed whether a coded bill was pro- or anti-industry (excluding those coded as Unknown) on an indicator for whether an insurance professional was involved in its drafting (as a sponsor or cosponsor or chair/member of a sponsoring committee).

The second added controls that captured the characteristics of the sponsors of each insurance bill. Many bills had multiple sponsors and cosponsors, of course; as a simple test, we collected data on each bill’s primary sponsor²⁴ (an approach we felt was more defensible than averaging the characteristics of sponsors and cosponsors). Our analysis included controls for the characteristics of sponsors’ districts that might be relevant to insurance regulation, including district per capita income (in $10,000s), the percentage of the district population employed in the finance or insurance industries, and the percentage of the district without health insurance.²⁵ We also controlled for the personal characteristics of each bill’s main sponsor: party affiliation, gender, race/ethnicity (measured simply as whether the member was white and non-Hispanic), and years in office. Of course, some of these controls may themselves be driven by our explanatory variable—for example, insurers may tend to partner with (or be) Republicans more often because of the party’s more lax stance on government regulation. In that sense, this approach represents a conservative estimate of the difference between insurer-backed bills and other legislation.

Our findings are reported in Table 4. In our first model (without controls), insurer-backed bills were about six percentage points more likely to be pro-industry. In our second model (with controls), insurer-backed bills were three percentage points more likely to be pro-industry, although the gap was no longer statistically significant. In the third model, we interacted the occupation of the sponsor with their party, under the suspicion that Republican insurance professionals may be more likely to sponsor pro-industry insurance bills than Democratic insurance professionals. The model’s estimates were consistent with this expectation, but the relevant interaction term was not statistically significant.

Table 4.

What Predicts Pro-Industry Bills?

	(1)	(2)	(3)
Insurer-backed bill	0.29*	0.11	0.04
	(0.11)	(0.11)	(0.19)
Sponsor characteristics
Republican		0.39*	0.38*
Republican		(0.08)	(0.09)
Insurer-backed bill × Republican			0.11
Insurer-backed bill × Republican			(0.24)
District per capita income (in $10,000s)		−0.16*	−0.16*
District per capita income (in $10,000s)		(0.06)	(0.06)
% constituents employed in insurance or finance		−0.01	0.01
% constituents employed in insurance or finance		(0.02)	(0.02)
% constituents without health insurance		−0.00	−0.00
% constituents without health insurance		(0.01)	(0.01)
Female legislator		−0.17	−0.17^†
		(0.10)	(0.10)
White legislator		0.37*	0.38*
		(0.13)	(0.13)
Years in office		−0.00	−0.00
		(0.00)	(0.00)
Intercept	−0.33*	−0.33	−0.33
	(0.04)	(0.24)	(0.24)
N	2,886	2,839	2,839
BIC	3,947.89	3,872.85	3,880.59

Source. Authors’ data collection and American Community Survey. Standard errors in parentheses.

†

p < .10. * p < .05, two tailed.

Overall, it appears that the differences between insurer-backed insurance bills and other insurance bills are not as large as a full-blown industry capture model might predict. In the 2011–2012 legislative term, many of the bills insurers supported imposed new burdens on the industry, and many of the bills put forth by non-insurers served to help insurance companies. Moreover, the majority of the bills that insurers sponsored were not friendly to the industry. When it comes to agenda-setting, it appears that occupation is not destiny. But our findings suggest that occupation plays some role in agenda-setting behavior. The mix of insurance bills that were introduced by insurers were more favorable to the industry than the mix introduced by non-insurers. As industry organizations have long suspected, having your own people in office can help push the legislative agenda (even if only marginally) in your favor.

State Legislative Agendas and the Occupational Makeup of Government

Taken together, the evidence from these three studies suggests that lawmakers who worked in insurance may in fact alter the legislative agenda in ways that promote what insurers want. State legislatures with more insurance professionals in them tend to spend less time considering new insurance regulations. When states consider insurance bills, however, they are disproportionately introduced by lawmakers from insurance backgrounds. And while former insurers’ bills on the whole are not particularly friendly to the industry, they are somewhat more likely to be industry-friendly than the bills their colleagues introduce.

These findings make several key contributions to the existing research on legislators’ occupational backgrounds. First, they suggest that the occupational makeup of government can affect not just roll-call votes but the legislative agenda itself. They also illustrate how a shift from simply studying legislative effort to studying both legislative effort and legislative content can help to shed light on both positive and negative forms of agenda setting.

When it comes to occupations and agenda-setting, negative and positive agenda-setting go hand in hand. For most sectors of the economy, government regulation is often burdensome, and negative agenda control is often the goal. But favorable legislation is always attractive to every industry. To the extent that a legislator’s occupation or industry affects his or her agenda-setting behaviors, we might expect to see both negative and positive forms of agenda control, as lawmakers reflecting their industry’s perspective simultaneously seek to minimize regulation while pushing for a few plum proposals.

More broadly, these findings underscore the importance of legislators’ occupational backgrounds. Political scientists have written a great deal about the difference women make in office, and the difference racial and ethnic minorities make. Lawmakers from different occupational backgrounds seem to make a difference, too. Our analysis of state legislative agendas suggests that there are indeed notable (albeit small) aggregate-level relationships between the numerical or descriptive representation of different professions and the legislative agendas in American states. The occupational makeup of the people the voters elect may shape one of the most important aspects of the legislative process: the problems and policies that make it onto the agenda.

Of course, this study has a number of limitations that are worth reiterating here. We only have data on three terms worth of legislative agendas, and we only have data from 30 states. The controls in Study 3 cover only primary sponsors, not all legislators involved in each bill’s introduction. We have lumped insurance professionals together, setting aside for now potential differences between insurers working in different areas (like health insurance vs. home insurance). And, most importantly, we have focused on just one industry. We have speculated that politicians from other large, well-organized, government-focused industries like finance, agriculture, law, and health care might have similar impacts on the legislative process, but follow-up research and replications will be needed to see whether these findings generalize to lawmakers from other occupational backgrounds.

It is also worth reiterating that the associations we document sometimes drop below conventional levels of statistical significance, especially in auxiliary models. Models with and without controls find significant negative associations between the numbers of insurance professionals in a legislature and the amount of new legislation affecting insurance, as do auxiliary models that drop outliers (Online Appendix Table A3), test alternative model specifications (Online Appendix Table A3), and add lagged dependent variables to the list of controls (Online Appendix Table A4). However, the negative association between insurance professionals and new insurance legislation falls short of statistical significance in a model with some but not all of our controls (Table 1; Online Appendix Table A4), three models estimated with data from just a single year (Online Appendix Table A2; Online Appendix Table A5), and one model estimated using an alternative OLS model specification (Online Appendix Table A3). Likewise, we find that insurance professionals introduce insurance bills that are significantly more favorable to industry, but when we control for partisanship and district characteristics, this association falls short of statistical significance (Table 3). On balance, the weight of the evidence is consistent with the idea that insurance professionals discourage industry regulation, play an oversized role in insurance policymaking, and tend to promote more insurance reforms that are favorable to industry. But these nonsignificant findings suggest that additional research is still needed. Data on the occupational backgrounds of politicians are still scarce—especially in state-level analyses, where we are often limited to samples of under 100. With larger samples, future researchers could more confidently assess the robustness of the findings in this analysis.

Even so, the results reported here are broadly consistent with past research and point clearly to the conclusion that occupations matter. Just as the number of women or people of color in public office affects whether gender and racial issues make it onto the agenda, the number of people from different occupations may affect whose voice is heard in our political institutions—as industry organizations have long suspected.

Supplemental Material

Insurance_-_SPPQ_Appendix_with_Data_Verification – Supplemental material for What Happens When Insurers Make Insurance Laws? State Legislative Agendas and the Occupational Makeup of Government

Supplemental material, Insurance_-_SPPQ_Appendix_with_Data_Verification for What Happens When Insurers Make Insurance Laws? State Legislative Agendas and the Occupational Makeup of Government by Eric R. Hansen, Nicholas Carnes and Virginia Gray in State Politics & Policy Quarterly

Footnotes

Acknowledgements

The authors thank Nathaniel Birkhead, Anthony DiMaggio, James Monogan, Marjorie Sarbaugh-Thompson, Kelsey Shoub, and participants in the 2017 Midwest Political Science Association Annual Meeting and the 2017 State Politics and Policy Conference for their helpful comments. The authors are grateful for research assistance from Tim Dinehart, Mason Dufresne, Dan Gustafson, David Joyner, Melissa Lee, Amanda Lewellyn, Katie Pischke, and Adam Weber.

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

Supplemental Material

Supplemental material for this article is available online.

ORCID iD

Eric R. Hansen

Author Biographies

Eric R. Hansen is Assistant Professor of political science at Loyola University Chicago. His research interests include legislative politics, state politics, social class, and representation.

Nicholas Carnes is the Creed C. Black Associate Professor of Public Policy and Political Science at Duke University. His research focuses on the economic and social class backgrounds of politicians.

Virginia Gray is the Robert Watson Winston Distinguished Professor of Political Science at the University of North Carolina, Chapel Hill. She has published widely on the topics of interest groups, state politics, and public policy.

References

Bachrach

Peter

Baratz

Morton S.

1962. “Two Faces of Power.” American Political Science Review 56 (4): 947–52.

Battista

James Coleman

. 2012. “State Legislative Committees and Economic Connections: Expertise and Industry Service.” State Politics & Policy Quarterly 12 (3): 284–302.

Battista

James C.

2013. “Financial Interests and Economic Diversity in State Legislatures.” Social Science Quarterly 94 (1): 175–99.

Baumgartner

Frank

Berry

Jeffrey M.

Hojnacki

Marie

Kimball

David C.

Leech

Beth L.

2009. Lobbying and Policy Change: Who Wins, Who Loses, and Why. Chicago: University of Chicago Press.

Bellemare

Marc F.

Carnes

Nicholas

. 2015. “Why Do Members of Congress Support Agricultural Protection?” Food Policy 50 (1): 20–34.

Berkman

Michael B.

O’Connor

Robert E.

1993. “Do Women Legislators Matter? Female Legislators and State Abortion Policy.” American Politics Quarterly 21 (1): 102–24.

Berry

William D.

Ringquist

Evan J.

Fording

Richard C.

Hanson

Russell L.

1998. “Measuring Citizen and Government Ideology in the American States, 1960-93.” American Journal of Political Science 42 (1): 327–48.

Bonica

Adam

. 2017. “Why Are There So Many Lawyers in Congress? ” Presented at the annual meeting of the Midwest Political Science Association, Chicago, April 6–9.

Bratton

Kathleen A.

Haynie

Kerry L.

1999. “Agenda Setting and Legislative Success in State Legislatures: The Effects of Gender and Race.” Journal of Politics 61 (3): 658–79.

10.

Broockman

David E.

2013. “Black Politicians Are More Intrinsically Motivated to Advance Blacks’ Interests: A Field Experiment Manipulating Political Incentives.” American Journal of Political Science 57 (3): 521-36.

11.

Buchanan

William

. 1962. “Subject Matter Experts.” In The Legislative System: Explorations in Legislative Behavior, eds. Wahlke

John C.

Eulau

Heinz

Buchanan

William

Ferguson

Leroy C.

, 193-215. New York: Wiley & Sons.

12.

Burden

Barry C.

2007. The Personal Roots of Representation. Princeton: Princeton University Press.

13.

Burstein

Paul

. 2014. American Public Opinion, Advocacy, and Policy in Congress: What the Public Wants and What It Gets. New York: Cambridge University Press.

14.

Campbell

Rosie

Cowley

Philip

. 2014. “Rich Man, Poor Man, Politician Man: Wealth Effects in a Candidate Biography Survey Experiment.” British Journal of Politics and International Relations 16 (1): 56-74.

15.

Canon

David T.

1999. Race, Redistricting and Representation: The Unintended Consequences of Black Majority Districts. Chicago: University of Chicago Press.

16.

Carnes

Nicholas

. 2012. “Does the Numerical Underrepresentation of the Working Class in Congress Matter?” Legislative Studies Quarterly 37 (1): 5-34

17.

Carnes

Nicholas

. 2013. White-Collar Government: The Hidden Role of Class in Economic Policy Making. Chicago: University of Chicago Press.

18.

Carnes

Nicholas

. 2016. “Who Votes for Inequality?” In Congress and Policy Making in the 21^St Century, eds. Jenkins

Jeffrey A.

Patashnik

Eric M.

, 106-34. New York: Cambridge University Press.

19.

Carnes

Nicholas

Hansen

Eric R.

2016. “Does Paying Politicians More Promote Economic Diversity in Legislatures?” American Political Science Review 110 (4): 699-716.

20.

Carnes

Nicholas

Lupu

Noam

. 2015. “Rethinking the Comparative Perspective on Class and Representation: Evidence from Latin America.” American Journal of Political Science 59 (1): 1–18.

21.

Carpenter

Daniel

. 2014. “Detecting and Measuring Capture.” In Preventing Regulatory Capture: Special Interest Influence and How to Limit It, eds. Carpenter

Daniel

Moss

David A.

, 57-68. New York: Cambridge University Press.

22.

Deaux

Kay

Reid

Anne

Mizrahi

Kim

Ethier

Kathleen A.

1995. “Parameters of Social Identity.” Journal of Personality and Social Psychology 68 (2): 280-91.

23.

Domhoff

G. William

. 1967. Who Rules America?. New York: Prentice Hall.

24.

Eulau

Heinz

Sprague

John D.

1964. Lawyers in Politics: A Study in Professional Convergence. Indianapolis: Bobbs-Merrill.

25.

Fenno

Richard F.

Jr . 1973. Congressmen in Committees. New York: Little, Brown.

26.

Gray

Virginia

Lowery

David

Fellowes

Matthew

Anderson

Jennifer

. 2005. “Legislative Agendas and Interest Advocacy: Understanding the Demand Side of Lobbying.” American Politics Research 33 (3): 404-34.

27.

Griffin

John D.

Anewalt-Remsburg

Claudia

. 2013. “Legislator Wealth and the Effort to Repeal the Estate Tax.” American Politics Research 41 (4): 599-622.

28.

Griffin

John D.

Newman

Brian

. 2008. Minority Report: Evaluating Political Equality in America. Chicago: University of Chicago Press.

29.

Grose

Christian R.

2011. Congress in Black and White: Race and Representation in Washington and at Home. New York: Cambridge University Press.

30.

Hall

Richard L.

1996. Participation in Congress. New Haven: Yale University Press.

31.

Hamm

Keith E.

Hedlund

Ronald D.

, and Stephanie Shirley Post. 2011. “Committee Specialization in U.S. State Legislatures during the 20th Century: Do Legislatures Tap the Talents of Their Members?” State Politics & Policy Quarterly 11 (3): 299–324.

32.

Hansen

Eric R.

Treul

Sarah A.

2015. “The Symbolic and Substantive Representation of LGB Americans in the U.S. House.” Journal of Politics 77 (4): 955–67.

33.

Hinkle

Steve

Brown

Rupert J.

1990. “Intergroup Comparisons and Social Identity: Some Links and Lacunae.” In Social Identity Theory: Constructive and Critical Advances, eds. Abrams

Dominic

Hogg

Michael A.

, 48-70. New York: Springer-Verlag.

34.

Jansa

Joshua M.

Gray

Virginia

. 2017. “Captured Development: Industry Influence and State Economic Development Subsidies in the Great Recession Era.” Economic Development Quarterly 31 (1): 50–64.

35.

Kingdon

John W.

[1984] 2011. Agendas, Alternatives, and Public Policies. Boston: Little, Brown.

36.

Koger

Gregory

. 2003. “Position Taking and Cosponsorship in the U.S. House.” Legislative Studies Quarterly 28 (2): 225–46.

37.

Kollman

Ken

. 1998. Outside Lobbying: Public Opinion and Interest Group Strategies. Princeton: Princeton University Press.

38.

Kraus

Michael W.

Callaghan

Bennett

. 2014. “Noblesse Oblige? Social Status and Economic Inequality Maintenance among Politicians.” PLoS ONE 9 (1): e85293.

39.

Krehbiel

Keith

. 1991. Information and Legislative Organization. Ann Arbor: University of Michigan Press.

40.

Kwak

James

. 2014. “Cultural Capture and the Financial Crisis.” In Preventing Regulatory Capture: Special Interest Influence and How to Limit It, eds. Carpenter

Daniel

Moss

David A.

, 71-98. New York: Cambridge University Press.

41.

Little

Thomas H.

Dunn

Dana

Deen

Rebecca E.

2001. “A View from the Top: Gender Differences in Legislative Priorities among State Legislative Leaders.” Women and Politics 22 (4): 29–50.

42.

Lowery

David

Gray

Virginia

Cluverius

John

Harden

Jeffrey J.

2012. “Explaining the Anomalous Growth of Public Sector Lobbying in the American States, 1997–2007.” Publius: The Journal of Federalism 43 (4): 580–99.

43.

Maddox

H. W. Jerome

. 2004. “Opportunity Costs and Outside Careers in U.S. State Legislatures.” Legislative Studies Quarterly 29 (4): 517–44.

44.

Matthews

Donald R.

1964. The Social Backgrounds of Political Decision-makers. New York: Random House.

45.

Mayhew

David R.

2000. America’s Congress: Actions in the Public Sphere. New Haven: Yale University Press.

46.

Millard

Sean

. 2015. “The US Insurance Industry.” Market Realist. http://marketrealist.com///us-insurance-industry-largest-world/201502 (accessed March 30, 2017).

47.

Miller

Mark C.

1995. The High Priests of American Politics: The Role of Lawyers in American Political Institutions. Knoxville: University of Tennessee Press.

48.

Minta

Michael D.

2011. Oversight: Representing the Interests of Blacks and Latinos in Congress. Princeton: Princeton University Press.

49.

Pitkin

Hannah Fenichel

. 1967. The Concept of Representation. Berkeley: University of California Press.

50.

Putnam

Robert D.

1976. The Comparative Study of Political Elites. Englewood Cliffs: Prentice-Hall.

51.

Renzulli

Dianne

and Center for Public Integrity. 2002. Capitol Offenders: How Private Interests Govern Our States. Washington, DC: Public Integrity Books.

52.

Schattschneider

E. E.

1960. The Semisovereign People. New York: Holt, Rinehart and Winston.

53.

Schiller

Wendy J.

1995. “Senators as Political Entrepreneurs: Using Bill Sponsorship to Shape Legislative Agendas.” American Journal of Political Science: 39:186–203.

54.

Sojourner

Aaron

. 2013. “Do Unions Promote Electoral Office-Holding? Evidence from Correlates of State Legislatures’ Occupational Shares.” Industrial and Labor Relations Review 66 (2): 467–86.

55.

Swers

Michele

. 2002. The Difference Women Make. Chicago: University of Chicago Press.

56.

Tajfel

Henri

Turner

John C.

1986. “The Social Identity Theory of Intergroup Behavior.” In Psychology of Intergroup Relations, eds. Worchel

Steven

Austin

William G.

, 7-24. Chicago: Nelson Hall.

57.

Thomas

Sue

. 1991. “The Impact of Women on State Legislative Policies.” Journal of Politics 53:958–76.

58.

Tibshirani

Robert

. 1996. “Regression Shrinkage and Selection via the Lasso.” Journal of the Royal Statistical Society, Series B (Methodological) 58 (1): 267–88.

59.

Wawro

Gregory J.

2000. Legislative Entrepreneurship in the U.S. House of Representatives. Ann Arbor: University of Michigan Press.

60.

Weingast

Barry R.

Marshall

William J.

1988. “The Industrial Organization of Congress; or, Why Legislatures, Like Firms, Are Not Organized as Markets.” Journal of Political Economy 96:132–63.

61.

Whitby

Kenny J.

1997. The Color of Representation: Congressional Behavior and Black Interests. Ann Arbor: University of Michigan Press.

62.

Witko

Christopher

Friedman

Sally

. 2008. “Business Backgrounds and Congressional Behavior.” Congress & the Presidency 35:71–86.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.22 MB