This file is also available in Adobe Acrobat PDF format
Personal Roots of Representation
Rick Santorum and Arlen Specter were something of an odd couple when it comes to Washington politics. Although both were United States senators representing Pennsylvania, and both are Republican, their policy views could not have been more divergent. One sought compromise and generally avoided hot-button issues; the other was defined by such issues. Santorum was chair of the Republican conference in the Senate and an advocate of conservative policies, particularly on sensitive social issues such as abortion and gay marriage. Specter is pro-choice and often bucks his party, to the point of generating backlash from Republican colleagues. These fundamental differences in their preferences affected not only their voting records but their decisions about which issues to tackle.
Particularly on domestic social issues, it is difficult to overstate the differences that separate the two senators. As Congressional Quarterly put it, “An unalloyed social conservative, Santorum saves his toughest fights for the issue of abortion. He is a leader in the fight to outlaw a procedure that opponents call ‘partial birth’ abortion, which he has called ‘barbaric’ and ‘the calculated killing of the nearly born.’”1 Specter, in contrast, is one of the few true moderates left in Congress. Some of his Republican colleagues in the Senate even challenged his position as chair of the Judiciary Committee when he suggested that conservative judicial nominees would get serious scrutiny. CQ reports that, unlike Santorum, “Specter fights to soften anti-abortion language in the GOP platform, tries to outdo Democrats on education spending, and not infrequently votes with Democrats on issues such as extending unemployment benefits and allowing patients to sue their managed-care health plans.” Although a four-term senator, he barely managed to win the 2004 Republican primary election in which he faced a more conservative challenger.
As one summary indicator of their differences, consider Poole and Rosenthal’s measure of legislative ideology known as NOMINATE. Their procedure uses roll call votes to place legislators on a unidimensional scale where 1 is most liberal and 1 is most conservative. By this yardstick, Santorum came in at .44, making him the 15th most conservative senator in the 108th Congress. Specter was located at .06, ranking him 51st in a chamber of 100. Nearly every other roll call indicator shows the same gulf between the two, a gap that is larger than that between the political parties themselves.
Why were Santorum and Specter so different, not only in their positions but in their policy priorities? Unfortunately, the usual suspects that congressional scholars line up to explain policy differences do not provide much of an answer. Were they from different states, a constituency focus would be the place to start. But they shared a common constituency in Pennsylvania, so electoral explanations cannot be responsible for the differences. Pennsylvania politics tend to be bifurcated between the eastern and western ends of the state, so it is possible that Santorum and Specter divide up the state geographically.2 Yet the differences in their voting records are far greater than demographic differences between Pittsburgh and Philadelphia would dictate. Both Santorum and Specter were seasoned politicians with years of congressional experience (and law degrees), thus rendering career-based explanations unhelpful. Finally, and most obviously, both senators were Republicans, rendering a partisan explanation for their actions moot. If career, constituency, and party affiliation fail, what accounts for the stark differences between Santorum and Specter? In reviewing their cases, the importance of more personal factors becomes obvious.
As a start, consider some facts about Santorum’s background. His CQ biography documents, “The son of an Italian immigrant—a Veterans Administration psychologist—the boyish-faced Santorum stresses individual responsibility and self-reliance, not government assistance programs. He can draw on his own experience of holding down a full-time job while attending law school.” Knowing just this, it might not be especially surprising that Santorum is conservative on fiscal matters.
But Santorum is best known for his activism on social issues, primarily his opposition to abortion, gay marriage, and stem cell research. He is a devout Catholic who attends Mass daily3 and held a course on Catholicism for fellow Republican members of Congress.4 He even helped convert fellow senator Sam Brownback (R-KS) to Catholicism. A former aide went so far as to call him “a Catholic missionary who happens to be in the Senate.”5
Despite the obvious importance of Santorum’s faith, he did not immediately tackle difficult morality issues when first elected to Congress. He was conservative on social matters but not especially active. “Santorum was pro-life from the time he first ran for Congress, but it wasn’t an issue he was vocal about. Santorum never gave a speech about abortion in the House or Senate until the late-term-abortion debates of 1996 and 1997.”6 The timing of this shift from passive to active opposition corresponds to a critical event in Santorum’s personal life.
In 1996 Santorum’s wife, Karen, was due with their fourth child. The pregnancy suffered from complications, and doctors warned that the baby had a fatal defect. Twice the doctors proposed aborting the fetus, an idea that the Santorums rejected. To cope with the pregnancy’s uncertainty, Karen began writing letters to the unborn child, already named Gabriel Michael Santorum. As doctors warned, Gabriel was born prematurely at five months and lived only a couple of hours outside the womb. Rather than follow standard protocol, the Santorums took the body home rather than to a funeral home and changed a hospital form’s description from “20-week-old fetus” to “20-week-old baby.”7 Although understandably upset about the outcome, Santorum was nonetheless satisfied that Gabriel died “naturally” rather than by an abortive “murder.” Now the father of six children, Santorum holds memories of Gabriel close. He keeps a framed photo of the baby on his desk. His wife’s prenatal letters to Gabriel were published as a book that condemns “infanticide” and “partial-birth abortion” while advocating for her husband’s renewed fight against such practices in the Senate. Santorum believes that Gabriel’s death intensified his views about abortion, which he now sees as the great moral issue facing America. His intensity about the issue clearly stems from personal experience.
Arlen Specter’s values come from an entirely different orientation, but likewise have their roots in life experience. They start not with religion but with a prosecutorial nose for the facts. A former district attorney, Specter is known for his tenaciousness, strong work ethic, and especially his independence (Fenno 1991). His independent streak motivated him to switch parties early in his career to increase his chances of winning office. Specter is also the son of a Russian immigrant and one of just a few Jewish Republicans in Congress. When called “ambitious,” Specter “reaches back to his immigrant roots,” saying, “When I was a growing up ambition was not only a good word in our household, it was an indispensable word” (quoted in Fenno 1991, 5).
Specter was never a typical Republican on social issues, and became less so over time as the party moved rightward in the 1980s and 1990s. Being largely pro-choice in a strongly pro-life party is challenging, as is representing the many Catholic and other antiabortion voters in Pennsylvania. Immediately after his first Senate victory Specter realized, “My toughest issue is abortion.” But he maintained his moderate position because, he said, “I’m convinced I’m on the right side” (quoted in Fenno 1991, 28). Yet abortion is not a great moral imperative for Specter, and he generally avoids making much commotion around the issue. His floor votes on abortion-related matters are generally centrist, and speeches on the issue are nonexistent.
The same cannot be said for health care, where he is an activist driven by deep personal experience. It began in 1993 when Specter complained to doctors that he suffered from pain in his face and neck. He suspected a brain tumor was responsible and asked for an MRI (magnetic resonance imaging) scan. His doctors declined the request because he showed no signs of memory loss or other common symptoms. Only after demanding it was Specter given the MRI; the test that revealed a two-inch brain tumor.8 Fortunately, the tumor was benign and could be removed with surgery, but the experience had an indelible impact.
Specter returned to the Senate soon after the surgery with a new scar on his head and a new attitude toward national health care policy. He immediately became an advocate of access to medical screening technology and preventive medicine. As chair of the Appropriations subcommittee dealing with health, he held some of the first hearings on breast cancer screenings. Specter argued that mammograms ought to be available to all women under 40, noting his own experience, in which a brain scan helped find a life-threatening tumor.9 Close observers of the budget process believe that the National Institutes of Health (NIH) get more funding than they would without Specter’s support.10
In sharp contrast to Santorum, Specter was also an early proponent of embryonic stem cell research. He became an even stronger advocate about the time he was diagnosed with Hodgkin’s lymphoma, a cancerous condition that was treated with chemotherapy. Upon introducing another bill to support therapeutic cloning in early 2005, Specter explained that his condition might have been treated with stem cell therapies if they had been supported by federal policies. Referring to his obvious hair loss caused by chemotherapy, Specter stated, “I’ve got a new hairdo, which you can all observe, and that is indicative of a problem which may well be helped by stem cell research if it were to go forward.”11 While several other senators joined his effort, Specter attracted attention and credibility because of his firsthand experience. “I don’t choose to unduly personalize it, but I have to find some way to excuse my hairdo.” Aside from Orrin Hatch (R-UT), who also had personal reasons for supporting such research, few Republicans were willing to buck President Bush and their party on the issue.
The lesson to be taken from Santorum and Specter is the potential for representatives’ experiences, interests, and expertise to shape their actions, what I call the personal roots of representation. Although the usual suspects of partisanship and constituency preferences surely play a role in guiding representatives’ actions in general, those two factors are far from determinative, even on some of the most salient issues in contemporary American politics. Santorum’s and Specter’s legislative preferences on specific issues are more complex than the current political science formulation allows. More importantly, their differences are not limited to positions taken in roll call votes. They also choose to be active on different sets of issues, with Santorum pushing for limits on abortion and Specter supporting access to health care and aggressive stem cell research.
This book is about the preferences that drive all of the Rick Santorums and Arlen Specters who inhabit the modern Congress. Their cases demonstrate that legislative preferences cannot be boiled down to partisanship and constituent interests. Preferences depend on a wider range of phenomena, including the backgrounds of legislators themselves. The information gleaned, and the interests and values formed, from life experiences shape their behavior on roll call votes, and more importantly and frequently, their proactive leadership on a smaller set of issues.
Unpacking Legislative Preferences
What members of Congress do surely depends on what their constituents want. Representatives are generally in tune with their constituents when it comes to broad orientations toward public policy. As I elaborate in chapter 2, evidence shows that liberal districts tend to elect liberal legislators, and conservative districts typically have conservative representatives. The bird’s-eye view strongly suggests that constituents are mostly getting what they want, just as the “delegate” view of representation expects. Yet I contend that this highly aggregated approach to assessing representation misses important deviations from public opinion. Relying only on summary information culled from many districts and many issues often leads to the false conclusion that legislators are strictly controlled by their constituents. But legislators can deviate from their constituents’ preferences. Why might they be allowed to do so?
Elections are thought to be the main mechanism by which legislators are held accountable. Madison argued as much in the Federalist Papers, and scholarly wisdom has long held that an unrepresentative legislator will be defeated by one whose policies are closer to the voters. Building on the median voter theorem, Downs (1957) made that point about party politicians, and Mayhew (1974) applied the logic of pandering to members of the House. The fear of losing one’s seat ought to be enough for the representative to act on constituents’ desires.
While the power of elections to motivate politicians is quite real, there are two main reasons why they are not sufficient to guarantee full policy representation. First, voters are only permitted to choose candidates, not policies. A candidate represents a bundle of policies, many of which are inconsistent, not well formed, or unknown to voters (McCrone and Kuklinski 1979). With only two candidates to choose from in most elections, voters will not find a candidate with whom they agree on every issue. Many congressional elections instead revolve around a small number of highly salient issues at the expense of many other matters that are likely to be on the legislative agenda. The ballot is just too crude an instrument to send finely tuned messages on desired policies.
Second, most subpresidential elections in the United States are not seriously contested. Only a few dozen of the 435 House races held every two years are truly competitive in the sense that both candidates have a reasonable chance of winning (Jacobson 2004). About three out of every 20 House elections are not contested at all (Wrighton and Squire 1997). The spatial voting model, in which voters choose candidates nearest their positions, only applies when candidates are equally credible and equally able to communicate with voters. When candidates have advantages in name recognition, staffing, and funding, they can deviate from public opinion to some degree without jeopardizing their seats (Burden 2004a). Indeed, members of Congress intentionally cultivate their constituencies to generate trust and leeway in policy (Bianco 1994; Fenno 1978). Public opinion would have more substantial influence on governing behavior if legislators feared retribution for not heeding it, but warding off competition dampens the influence of constituents. Because elections are inadequate to ensure that voters’ views on policy are fully represented by candidates, legislators will never perfectly represent the policy wishes of their constituents.
In a principal-agent framework used in economics, the job of the constituent is to monitor what legislators do. As the principal, the median voter in the district employs an agent in Washington to act on her behalf. I have already argued that the principal faces a difficult task because neither fear of replacement nor strict monitoring will be sufficient to ensure the agent’s compliance. In this framework we should thus expect slack between constituents’ preferences and legislators’ actions, or what is known in much of the literature as “shirking.”12 When constituents’ views are not represented, it is usually assumed that the legislator is “shirking” his duties to pursue his own self-interest (Bender and Lott 1996; Uslaner 2002). But this reasoning jumps too quickly to a conclusion about the causes of slack. While identifying the degree of disagreement is a useful indicator of how much leeway a legislator has, the shirking approach says little about the causes of such slack. As I explain in chapter 2, it is not enough to estimate how much shirking there is; it is imperative to understand why it exists in the first place.
Portraying the situation as a contest between constituents’ preferences and the legislator’s ideology alone assumes a false dichotomy. Many factors shape legislative behavior. Among them are pressure from interest groups, party influence, cues from the administration, media priming, and logrolling among members. As a result, the standard “delegate” versus “trustee” paradigm that is so heavily used in legislative studies oversimplifies the task of understanding how representatives behave. The delegate model may be a useful standard of evaluation, since such a legislator simply follows her constituents’ wishes (assuming they can be identified). But if the trustee model is the only alternative to full policy representation, deciding which is correct unnecessarily distracts researchers from asking important questions about other influences.
To evaluate the health of republican government, it is critical for one to know the conditions under which a legislator heeds or dismisses the opinions of his constituency. Researchers must begin decomposing policy shirking into its constituent elements. To offer a medical analogy, a doctor cannot evaluate a person’s health knowing only how much cholesterol is in his blood. The type of cholesterol is just as important. In addition to other indicators, the physician needs to distinguish between high-density lipoproteins and low-density lipoproteins (HDL and LDL), otherwise known as “good” and “bad” cholesterol. Although bad cholesterol can clog artery walls, good cholesterol actually helps clear blood vessels. Likewise, the total amount of shirking or policy leeway is only properly understood when the reasons for it are known. The total amount of shirking is a function of both “good” and “bad” variables. In the final chapter of the book I will suggest which forms of shirking might be considered desirable and which undesirable.
This approach points the study of congressional representation in a new direction. I care not only where a politician stands in relation to her constituents, but why and with what consequences. Abstract ideological preferences can be measured reasonably well, but we must unpack them to understand how the member came to his positions and how they motivate legislative action. Two legislators might take positions contrary to their constituencies for different reasons. The first could do so for principled reasons. She might have chosen her position by reasoning from a broader philosophy of government or because of expertise that most constituents lack. The second legislator might have less noble motives. His party leaders pressured him to take the position, or an interest group’s lobbying was effective. Both members have shirked, but for quite different reasons.
Levels of legislative shirking are important too, perhaps more important in the early stages of the lawmaking process than the eventual positions themselves. On many issues, legislators merely vote on the relevant roll calls as the come, but on a small set of issues a legislator becomes much more heavily involved. Time is a precious resource, and most of a legislator’s time is spent on just a handful of issues. As players in particular domains of policy, they give speeches, influence colleagues, and sponsor legislation. If these issues happen to be the ones on which they differ from their constituents, the degree of deviation is more severe than it would be if only roll call votes were at stake. Proactive behavior by legislators upstream in the policy process largely determines the agenda that other legislators will vote on.
My perspective encourages a different empirical approach than is usually taken in the study of Congress. The typical study tries to explain general roll call voting patterns, lumping all members into a single analysis. When an explanatory variable such as party affiliation turns out to be a statistically significant predictor, the researcher concludes that partisanship causes legislators’ voting patterns. Although this can be a useful enterprise, I believe that it misses much of congressional reality.
There are three ways in which my approach diverges from this standard method. First, I examine activity on specific policies rather than general indicators of preferences or activity levels. The prototypic congressional study relies on NOMINATE scores (Poole and Rosenthal 1997) or ratings by interest groups as indicators of a legislator’s ideology. Which of these general measures is used does not matter much (Burden, Caldeira, and Groseclose 2000) since each one implies a similar view of representation. Each roll call scoring technique assumes that the member’s actions across a wide range of issues can be summarized by a single position. Roll call voting is often the only standard by which legislative activity is judged. Yet it would be a grave mistake to assume that every issue is defined on the left-right continuum just because such ideological scales can be constructed. This approach also assumes that legislative activities prior to the vote itself are inconsequential. In later chapters, I challenge both of these ideas. Exploding these assumptions makes an issue-by-issue approach more appropriate. Even after controlling for general ideological preferences and previous levels of support for a position, I am able to identify member-specific factors that drive legislators’ positions and their attention to issues. The experiences of Rick Santorum and Arlen Specter show that these personal factors can have more influence on congressional behavior than is commonly thought.
Second, my theory encourages analysis of legislators in the majority and minority separately. I argue that the effects of many of the member-specific variables are asymmetric across parties, depending on whether the party is a proponent or opponent of the policy status quo. Generally speaking, threat motivates more personal involvement than does opportunity. The tendency to throw all legislators into a single analysis hides these important distinctions. I demonstrate this point in several cases by first analyzing all representatives simultaneously, then separating by party to show the differential effects of members’ traits. I would reason, for example, that religious faith and personal health experiences affect Santorum and Specter much more than they would two equivalent Senate Democrats.
Third, and building on the first two points, I contrast proactive and reactive behavior. Proactive behavior is action that requires initiative. It includes doing research outside of committee, lobbying fellow members, contacting interest groups, making floor speeches, and introducing and cosponsoring legislation. Although the range of possible proactive behaviors is great, for tractability I focus on speechmaking and bill cosponsorship. What all proactivity has in common is that it requires a conscious decision by the legislator to take action in advance of a floor vote. As Hall (1993) has shown, members only have enough resources to be active on a handful of issues, and their decisions about which issues to take up are not made lightly. The factors that explain proactive behavior are quite consequential in the process of representation since proactive behavior magnifies a legislator’s influence. Reactive behavior, in contrast, is a passive response to decisions made by others. The quintessential reactive behavior is roll call voting. Because research on Congress focuses so heavily on roll calls, we know little about the forces that differentiate the two kinds of actions (Van Doren 1990). If constituents or personal traits have different effects on reactive and proactive behavior, as I argue they do, then the distinction is critical empirically, theoretically, and normatively. Santorum’s and Specter’s backgrounds are responsible for their differing roll call records across a wide range of bills, but they play an even more substantial role in motivating proactivity on a small set of issues.
As a corollary to these three innovations, I ask whether campaign contributors are aware of how legislators’ personal traits affect their behavior. In several cases I discover that patterns of giving by political action committees (PACs) are indeed sensitive to representatives’ backgrounds. PACs give more money to those with experiences sympathetic to their interests, even after accounting for partisanship, constituency preferences, and other factors. Not only do legislators’ personal roots shape their lawmaking activities, but they also affect how much campaign money can be raised, which in turn shapes their odds of reelection.
In addition to these three innovations, the book’s strength comes from its multimethod approach. I develop a theory based on a variety of literatures in political science and economics, but it is also informed by in-depth qualitative interviews with several former members of Congress. The interviews help ground the study and inform the hypotheses to be tested. I also conduct extensive quantitative analysis of legislative behavior, both reactive and proactive. In three different policy domains, I examine the propositions listed above and find support for the theory.
Descriptively, I have argued that policy representation is not perfect, in that constituents do not get exactly the policies they would like. From a normative standpoint, it could even argued that constituents should not get everything they want. Voters are less informed and deliberate less than legislators, yet this does not mean that legislators are always right. One might argue that the collective views of several hundred thousand constituents in a House district might effectively counter any biases contained in a single representative (Page and Shapiro 1992).13 The structure of Congress, with its committee-based system ofexpertise and moderately strong parties, does not always encourage policy responsiveness. The influence of party leaders, lobbyists, campaign activists, and the administration can pull members away from district concerns. These institutional factors have been studied at length. As important as these forces are, their effects are also inconsistent as one moves from issue to issue. New coalitions must be assembled on each piece of legislation. Party leaders are strong in one domain, but weak in another. Interest groups wield influence in one issue area but not others. More lasting are the traits of members themselves. Legislators acquire expertise, values, and interests long before they arrive in Washington. These early influences often lead to more intense preferences that have the potential to deter other influences, even pressure from constituents. The limits of policy representation are often found in the representative himself.
As I shall argue, voters would do better to elect someone who is like them than to hope to change a representative who is not. In the language of economics, a principal would be better served to find a like-minded agent who does not require monitoring than to attempt control a potentially disobedient agent. Elections are a more powerful mechanism for enhancing representation when constituents are able to choose a candidate who already shares their value systems. Although elected officials certainly monitor public opinion, it can be difficult for citizens to change legislators’ behavior on all but the most symbolic of issues. To the degree that the principal-agent framework adequately describes the relationship between the district and the representative, it must be enforced by selection at the outset, not conversion thereafter (Bernstein 1989). As a result, much of the policy representation that happens in Washington can be thought of as coincidental, at least in the sense that the representative happens to act in a way that constituents condone, rather than being controlled by voters.
The Rest of the Book
The remainder of the book explores the personal roots of representation. In chapter 2 I begin unpacking the concept of legislative “preferences” so often used by congressional scholars. In reviewing this literature, it seems ironic that most studies of policy responsiveness look at ideological behavior rather than policy itself. I develop a theory of legislator preferences that identifies why representatives position themselves as they do on particular issues, and what consequences these positions have. I propose values, information, and self-interest as three mechanisms that allow personal characteristics to be influential. The theory thus goes beyond taking generic “preferences” as mere starting points in explaining congressional activity by decomposing the sources of these preferences.
The dominant view among academics is that preferences are induced by constituents’ pressure. These preferences are mostly reactive and affect the tail end of the policy process—roll call votes—since they are best suited for taking a position. But this focus on roll call votes both overstates and understates the role of constituents in determining legislator behavior. Often a legislator will act in a way consistent with constituents’ preferences without purposely representing the district. The prevalence of coincidental factors means that analysis of roll call votes alone, particularly without measures of legislators’ interests, will overstate the degree of constituent control. Yet legislators can also enhance the influence of constituency wishes by being more active on a few issues of great concern to the people they represent.
I point to many cases where legislators, purposely and often actively, deviate from district opinion. In these situations their preferences are induced by other forces, some of which emanate from within the legislator herself. My interviews with six former legislators, coupled with a series of examples from the media, highlight the importance of appreciating the complexity of legislative preferences.
The next three chapters each apply the theory to a different policy area. The most comprehensive of the applications, chapter 3 analyzes policy-making around tobacco and smoking. I begin by describing the “tobacco wars” of the 1990s to establish the context in which defenders of tobacco were especially threatened by policy change on the way. I estimate regression models that explain why legislators take pro- or antitobacco control actions in the 104th House. In addition to standard controls such as legislators’ ideological predispositions, partisanship, and district interests in the tobacco industry, I find that representatives’ personal tobacco use exerts direct and independent effects on their behavior. As defenders of the status quo, members who smoke are more likely to vote against, speak in opposition to, and sponsor bills to limit tobacco control measures ranging from advertising and sales prohibitions to ending of tobacco farm subsidies. Efforts in favor of tobacco control, in contrast, appear largely unrelated to personal and district ties to the industry and are more entrepreneurial in nature. Moreover, tobacco PACs appear to have identified sympathetic members of Congress and provide them with significantly more contributions than are given to other members with similar partisan or ideological predispositions.
Chapter 4 examines recent education policymaking, specifically vouchers and other school choice plans. Proposed reforms in the 1990s offered school choice and threatened the educational status quo of public education. I analyze action on a voucher initiative and more general voting patterns on education policy in the 107th House. Although school choice efforts have a strong partisan basis—with Republicans more likely than Democrats to favor alternatives to public schools—members’ actions are also products of their districts’ educational needs and their own experiences with school systems. Members with school-age children are more like to be active on the issue. Perhaps more importantly, legislators whose children are in public school are less likely to vote for a school voucher program. In addition, the relationships for threatened Democrats appear more conditional on the confluence of district and personal interests, again pointing to the imbalance in participation rates between the two sides of the issue. Finally, I show that teacher union PACs were sensitive to the school choices made by representatives. Legislators who sent their children to public school were more likely to receive contributions, particularly so if they were Democrats.
Chapter 5 analyzes a class of new ethical issues that include stem cell research, tax incentives for religious charities, and legal protection of religious practices. In each of these three cases I demonstrate that legislators’ religious faiths affected their behavior. The personal influences were greatest among Republicans who were most defensive of the status quo at the time and in early proactive stages of the lawmaking process, where participation is selective. In addition, I find that members of smaller religious denominations who faced greater threats were especially motivated to support the religious liberties protection bill.
The book concludes in chapter 6 by reviewing the theoretical argument and summarizing the findings across the three policy applications. I point to ways in which the conventional understanding of congressional policymaking has missed important relationships that govern the connections between legislators and their constituents. Elections are not especially efficient means for constituent influence over policy unless they result in the selection of like-minded representatives. While the electoral connection is often powerful on a select set of issues where constituents’ concerns are most salient, the nature of congressional deliberation allows for a great deal of policy shirking, both reactive and proactive. I also evaluate the normative value of deviation from constituency preferences using my theory as a guide. Determining the conditions under which legislators disregard constituents’ wishes provides a new perspective on policy representation in the United States, one that appreciates the fundamental motivations that cause elected officials to do what they do.
Return to Book Description
File created: 8/7/2007
Questions and comments to: email@example.com
Princeton University Press