SEARCH

SEARCH BY CITATION

Keywords:

  • democratic representation;
  • campaign promises;
  • roll-call votes;
  • environmental policy;
  • Project Vote Smart;
  • NPAT

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Campaign Promises, Representation, and Environmental Policy
  5. Connecting Campaign Promises to Policy Choices
  6. How Often Do Legislators Defect from Campaign Promises?
  7. Predicting Defection from Campaign Promises
  8. Results and Discussion
  9. Campaign Promises, SMDs, and Environmental Policy
  10. Concluding Remarks
  11. References
  12. Appendix

One important criterion for assessing the quality of democratic governance is the extent to which the policy process effectively translates citizen preferences into collective choices. Several scholars have observed a discrepancy between citizen preferences for strong environmental protection and weak policies adopted in the United States, indicating that the United States may fall short on this criterion. We examine one possible mechanism contributing to this discrepancy—legislator defection from campaign promises. Our data indicate that legislators in the U.S. Congress routinely defect from their campaign promises in environmental protection, undermining the link between citizen preferences and policy choice. We also find that legislators are much more likely to defect from pro-environmental campaign promises, which moves government policy toward less stringent environmental programs. Finally, the propensity of legislators to defect from their campaign promises is systematic, with defection affected by partisanship, constituency influence, the influence of the majority party, and the likely consequences of defection for policy choice. These findings contribute empirical evidence relevant to the “mandate theory” perspective on how citizen preferences are translated into collective choices through the policy process. These findings may also complement research in comparative politics concluding that legislatures selected through single member districts adopt less stringent environmental policies than do legislatures chosen via proportional representation in that the mechanism for this effect may go through legislator defection from campaign promises.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Campaign Promises, Representation, and Environmental Policy
  5. Connecting Campaign Promises to Policy Choices
  6. How Often Do Legislators Defect from Campaign Promises?
  7. Predicting Defection from Campaign Promises
  8. Results and Discussion
  9. Campaign Promises, SMDs, and Environmental Policy
  10. Concluding Remarks
  11. References
  12. Appendix

At the beginning of the modern environmental era, the United States was broadly looked to as a leader for the strength and scope of its environmental protection policies. More recently, observers have pointed out that environmental policies in the United States lag behind those in other advanced industrial democracies (Switzer, 2001; Vig & Faure, 2004). If citizen support for environmental protection is systematically lower in the United States, then less aggressive policies would be an indicator of effective democratic governance. But by most measures, citizen support for environmental protection is no different in the United States than in these other nations. For example, an average of 61.8 percent of citizens in 10 Western European nations and Japan responded that they would be willing to pay higher prices in return for stronger environmental protection regulations. In the United States during the same period, this figure is 65 percent. Similarly, an average of 64.7 percent of citizens in 10 Western European nations and Japan responded that they would support greater efforts at environmental protection even at the expense of economic growth. In the United States during the same time period, this figure is 58 percent (Dunlap, Gallup, & Gallup, 1999). An average of 96.8 percent of citizens in 17 Western European nations and Japan supported government efforts to protect the environment. In the United States during this same period, the figure is 96 percent (Mertig & Dunlap, 2001). Finally, an average of 8.4 percent of citizens in 17 Western European nations and Japan belonged to an environmental interest group. In the United States during this same period, the figure was 15.9 percent (Dalton, 2005). These observations suggest two research questions:

  1. Why does the policy process in the United States produce weaker environmental policies than citizens might prefer?
  2. Why does the policy process in the United States produce weaker environmental policies than in other industrial democracies?

In this manuscript we address the first question by examining one possible source of slippage between citizen preferences and public policy—the extent to which elected officials defect from their campaign promises. We investigate this mechanism by examining the correspondence between the environmental campaign promises and the policy choices made by members of the U.S. House of Representatives and U.S. Senate between 1992 and 2004. Specifically, we measure the extent of defection from campaign promises, and we predict the propensity of legislators to defect from their campaign promises. Our results indicate that legislators in the United States routinely defect from their campaign promises, undermining the link between citizen preferences and government policy choices, and posing difficulties for mandate theory as an explanation for how these preferences are translated into policy choices within democratic systems. In addition, we find that legislators are significantly more likely to defect from pro-environmental campaign promises, which has the effect of moving environmental policy toward less stringent programs. Our findings may also have implications for the second question, as more frequent defection from campaign promises in single-member district (SMD) legislative systems like the United States may help account for the observation that the United States and other SMD systems adopt weaker environmental policies than do nations with legislatures chosen via proportional representation (PR).

The remainder of the manuscript proceeds as follows. First, we examine the role of campaign promises in effective democratic governance. Second, we discuss our methods of measuring campaign promises and for linking campaign promises to legislator policy choice. Third, we look at the frequency with which members of the U.S. Congress defect from their campaign promises in environmental policy. Fourth, we describe the models used to predict the propensity of legislators to defect from campaign promises in environmental protection. Finally, we explore the implications of our results for the observation that legislatures selected through SMDs adopt less stringent environmental policies than do legislatures chosen via PR.

Campaign Promises, Representation, and Environmental Policy

  1. Top of page
  2. Abstract
  3. Introduction
  4. Campaign Promises, Representation, and Environmental Policy
  5. Connecting Campaign Promises to Policy Choices
  6. How Often Do Legislators Defect from Campaign Promises?
  7. Predicting Defection from Campaign Promises
  8. Results and Discussion
  9. Campaign Promises, SMDs, and Environmental Policy
  10. Concluding Remarks
  11. References
  12. Appendix

Campaign Promises and Effective Democratic Representation

The most common theory of democratic representation, often labeled mandate theory, sees popular preferences being translated into public policy through a majority of voters choosing representatives whose policies they prefer.1 Sullivan and O'Connor (1972) offer four conditions that must be met if elections are to facilitate public influence over policymaking via mandate theory:

  1. Opposing candidates for office must offer voters differing issue positions.
  2. Voters must perceive the issue positions of candidates.
  3. Voters must cast their ballots on the basis of these perceived issue positions.
  4. Winning candidates must vote in accordance with their preelection issue positions.

The requirements of mandate theory as employed by Sullivan and O'Connor (1972), McDonald, Mendes, and Budge (2004), and others are virtually identical to Mansbridge's notion of “promisory representation” where candidates seek to attract voters by making promises regarding future policy choices, and voters evaluate candidates based upon these promises (Mansbridge, 2003). In promisory representation, “voters' power works forward to hold representatives to the promises they made at election time” (Disch, 2011, p. 101). For both mandate theory and promisory representation, then, legislator fidelity to campaign promises is a necessary condition for effective democratic governance.

While there is widespread popular and scholarly skepticism regarding each of the four elements of mandate theory, the empirical evidence suggests that conditions 1–3 may be commonly met in practice, at least in the United States. First, the Democratic and Republican parties offer meaningfully different issue positions to voters both generally (Ansolabehere, Snyder, & Stewart, 2001a; Erikson & Wright, 2001; Kahn & Kenney, 2001) and regarding environmental policy in particular (Shipan & Lowry, 2001). Second, despite the perception that voters in Congressional elections are poorly informed, several studies indicate that many voters, much of the time, correctly perceive the issue positions of candidates (Aldrich, Sullivan, & Borgida, 1989; Ansolabehere & Jones, 2010; Buttice & Stone, 2012; Kahn & Kenney, 2001; Lodge, Steenbergen, & Brau, 1995; Wright & Berkman, 1986). Third, much of this research also concludes that voters' choices in elections are influenced by and/or reflect these perceived differences in candidate issue positions (Ansolabehere & Jones, 2010; Lodge et al., 1995; Toms & Van Houweling, 2008; Wright & Berkman, 1986). To quote Phillip Edward Jones, “the buck that stops with members of Congress is for the positions that they take, not for the policy outcomes that they preside over” (Jones, 2011, p. 764).

This evidence regarding the applicability of conditions 1–3 does not mean that the debate over these conditions is settled. This evidence does suggest, however, that citizens are more competent than previously believed, and therefore more capable of expressing coherent preferences and holding their representatives accountable for actions that are incompatible with these preferences (Disch, 2011). By contrast, we know very little about condition 4. One piece of early research demonstrated that members of the House of Representatives in the 90th Congress voted congruently with preferences expressed prior to their election (Sullivan & O'Connor, 1972). More recently, Ansolabehere, Snyder, and Stewart (2001b) found a high level of congruence between candidate policy positions and roll call votes in the 103rd through 105th Congress across several issue areas. By examining this question in the aggregate, however, neither of these studies evaluated the extent to which representatives defected from specific campaign promises. Ringquist and Dasse (2004) provide what may be the only systematic investigation of whether candidates for Congress keep specific campaign promises after they are elected. While these authors found that legislators voted consistent with their campaign promises in a majority of instances, this study is quite limited in that it examined a single chamber (the U.S. House of Representatives) for a single Congress (the 105th). More evidence is needed as to whether condition 4 of mandate theory holds in practice.

Campaign Promises, Environmental Politics, and Environmental Policy

The link between fidelity to campaign promises and effective democratic representation applies to all policy areas. We argue here that the characteristics of environmental politics make defection from campaign promises especially likely in environmental policy, and that these defections should move legislative choice toward less stringent environmental policies.

In the United States, environmental protection is a salient public issue but a much less salient electoral issue for the majority of citizens. That is, large majorities support greater governmental efforts to protect and improve environmental quality, but few voters cast their votes on the basis of environmental issues (see Dunlap, 1995; Rosenbaum, 2010). Moreover, opponents of environmental regulation generally have more resources, and carry more clout in Congress, than do supporters of environmental protection (Kamieniecki, 2006; Kraft, 2006). In this situation, candidates face strong incentives to make pro-environmental campaign promises consistent with public opinion. After the election, however, lobbyists and other powerbrokers employ their resources to encourage members of Congress to vote the anti-environmental position on legislation. Because few voters cast their ballots on the basis of the environmental policy choices of legislators, there are few electoral repercussions for this behavior. Such a situation is ripe for defection.

The argument in the preceding paragraph has two observable implications: first, that defection from campaign promises should be more common in environmental protection than in other policy areas; and second, that legislators should be more likely to defect from pro-environmental campaign promises than from anti-environmental campaign promises. This second implication might even be considered rational, as pro-environmental promises are consistent with general citizen preferences while defection from these promises is consistent with the preferences of the most powerful stakeholders in the legislative environment. A predominance of defection from pro-environmental campaign promises will produce more anti-environmental votes on legislation, thereby moving policy choices toward less stringent environmental policies. Because we only examine environmental policy, we cannot test the first implication. We can test the second.

Connecting Campaign Promises to Policy Choices

  1. Top of page
  2. Abstract
  3. Introduction
  4. Campaign Promises, Representation, and Environmental Policy
  5. Connecting Campaign Promises to Policy Choices
  6. How Often Do Legislators Defect from Campaign Promises?
  7. Predicting Defection from Campaign Promises
  8. Results and Discussion
  9. Campaign Promises, SMDs, and Environmental Policy
  10. Concluding Remarks
  11. References
  12. Appendix

In order to measure the frequency of defection from campaign promises, we need a measure of candidate campaign promises and a method for connecting these promises to specific policy decisions made by legislators.

Measuring Campaign Promises

No source of information regarding campaign promises is without problems. Relying upon campaign literature, advertising, or stories in newspapers presents serious logistical difficulties if one wants to obtain data for all major party candidates (though see Hill, 2001, and Sulkin, 2005). These sources pose two additional problems for systematic assessments of the extent to which candidates keep their campaign promises: (i) candidates may not express issue positions in the same policy areas, and (ii) the information these sources provide regarding candidate policy positions is often vague and typically avoids controversial issues (Klotz, 1997; West, 1993). Faced with these problems, previous scholars have gathered information on the preelection policy preferences of candidates through surveys (Barrett & Cook, 1991; Sullivan & O'Connor, 1972; Wright & Berkman, 1986). While these surveys overcome the problems noted previously, they cannot properly be considered campaign promises because the policy positions expressed in the surveys were never made public, and therefore were unavailable to voters.

We obtain data on candidate campaign promises from the National Political Awareness Test (NPAT). In each election since 1992, Project Vote Smart, a nonprofit, nonpartisan organization, has surveyed American legislative candidates about their positions on specific issues. The NPAT provides candidates with lists of policy alternatives, and asks the candidates to identify which of these options they support. We take our data for campaign promises from the sections of the 1992–2002 NPATs dealing with environmental protection.2 We coded each candidate's response to each of the 57 environmental policy questions asked during this period, attaching a score of 0 to “anti-environmental” responses (e.g., opposition to strengthening the Clean Water Act) and a score of 1 to “pro-environmental” responses (e.g., supportive of strengthening the Clean Water Act).3

We contend that this measure of the expression of candidates' intentions is superior to those used in the past. Critically, Project Vote Smart intends for NPAT responses to be interpreted as campaign promises, and advertises them as such (G. C. Wright, personal communication, January 15, 2006). Compared with the survey results used in previous research, NPAT responses are eminently public. News programs on all major networks, each of the major weekly newsmagazines, and dozens of the most prominent daily newspapers in the United States have featured articles on the candidate information contained in the NPAT. The NPAT website receives more than 16 million hits each day, and Project Vote Smart staff responds to hundreds of thousands of additional inquiries each year via telephone (Project Vote Smart, ). In some ways, information regarding candidate policy intentions from the NPAT may be of higher quality than similar information from the candidates' campaigns, as the NPAT forces all candidates to take clear positions on specific policy issues. A final indicator of the value of the NPAT is the frequency with which scholars are relying upon this survey to measure candidate issue positions (Ansolabehere et al., 2001a, 2001b; Erikson & Wright, 2001; Ringquist & Dasse, 2004; Shor, Berry, & McCarty, 2010; Shor & McCarty, 2011).4

Matching Campaign Promises to Legislator Policy Choices

We measure legislators' policy choices using roll call votes on environmental legislation. Roll calls do not represent all actions Senators and Representatives can take to affect public policy (Hall, 1996; Sulkin, 2005). Moreover, as legislation often contains elements relating to widely different policy areas, it is sometimes difficult to use individual roll call votes to gauge the specific policy consequences of actions taken by members of Congress. Most important legislative policy decisions, however, are recorded using roll call votes, and roll calls are the most commonly used measure of Congressional behavior in academic research. We take steps to assure that our roll calls represent specific policy decisions by excluding appropriations bills and omnibus bills. In addition, a large percentage of our roll calls are not on bills per se, but on specific amendments to these bills.

We employed three tactics for identifying roll call votes related to specific NPAT questions. First, we examined the brief reported summaries of all roll call votes taken in the House and Senate in the 103rd–108th Congresses. When these summaries indicated that a bill or amendment might be related to one of the NPAT questions, we obtained a detailed bill summary to confirm this possibility. Second, recognizing that the first approach might miss some relevant legislation, we crafted three keyword phrases for each of the 57 NPAT questions, used these keywords in searches of all bills and amendments introduced in the House and Senate in the 103rd–108th Congresses, and examined closely the summaries and legislative histories of all bills and amendments identified this way. Third, we compared all of the key roll call votes identified by the League of Conservation Voters (LCV) in the 103rd–108th Congresses with the 57 NPAT questions. Any relevant key votes that were not identified using the first two search protocols were added to the data set. Using these tactics, we identified 87 roll call votes in the House and 62 roll call votes in the Senate directly related to questions from the environmental policy sections of the NPAT. We coded legislators' votes on each roll call, assigning a value of 0 to anti-environmental votes and a value of 1 to pro-environmental votes.

In many cases, matching roll call votes to NPAT responses was quite easy. For example, the 2002 NPAT asked candidates whether they supported the increased use of alternative fuel technology. One matched roll call vote is 108.2.321 on House amendment 624 to increase spending on renewable energy programs by $30 million. In other cases, matches were more difficult. For example, the 1996 NPAT asked candidates for Congress whether they supported requiring the federal government to reimburse citizens when environmental regulations limit the use of private property. One matched roll call vote is 105.2.197 on Senate bill 2271, which simplified access to the federal courts for injured parties whose rights and privileges have been deprived by final actions of federal agencies. While the title and brief description of the bill are not obviously related to the NPAT question, a detailed examination of the bill's contents shows that the sole rationale for the bill was to make it easier for citizens to sue federal agencies for reimbursement when environmental regulations reduced the value or otherwise limited the use of their property. Clearly, then, a vote in favor of Senate bill 2271 is consistent with an NPAT response favoring governmental compensation for regulatory takings in environmental protection.

How Often Do Legislators Defect from Campaign Promises?

  1. Top of page
  2. Abstract
  3. Introduction
  4. Campaign Promises, Representation, and Environmental Policy
  5. Connecting Campaign Promises to Policy Choices
  6. How Often Do Legislators Defect from Campaign Promises?
  7. Predicting Defection from Campaign Promises
  8. Results and Discussion
  9. Campaign Promises, SMDs, and Environmental Policy
  10. Concluding Remarks
  11. References
  12. Appendix

We use the NPAT responses and matched roll call votes described previously to construct a dichotomous variable coded 1 whenever a member of Congress defects from their campaign promise (i.e., when their roll call vote is inconsistent with their NPAT response) and 0 otherwise. Aggregating this variable across all legislators and bills, we find that from 1993 through 2004, members of the House of Representatives defected from their campaign promises 31 percent of the time, while Senators defected 41 percent of the time (see Table 1).5

Table 1. Proportion of the Time Members of Congress Defect from Campaign Promises on Environmental Policy Roll Call Votes, 103rd–108th Congress
 House Defection RateSenate Defection Rate
Overall defection rate0.310.41
By valence of campaign promise  
Pro-environmental promise0.400.44
Anti-environmental promise0.220.34
By Congress  
103rd0.340.33
104th0.280.27
105th0.340.34
106th0.340.52
107th0.470.45
108th0.210.41
By issue type  
Pollution control0.250.37
Natural resources0.270.34
Property rights0.280.23
Regulatory reform0.340.17
Climate change0.300.42
Energy0.420.45
Endangered species0.480.49
By party  
Democrat0.340.30
Republican0.290.47

These single figures for each chamber mask significant heterogeneity in defection. First, defection is much more likely from pro-environmental campaign promises than from anti-environmental campaign promises. In the House of Representatives, anti-environmental campaign promises were kept nearly 78 percent of the time (defection rates were 22.1 percent), while pro-environmental campaign promises were kept only 60.4 percent of the time (defection rates were 39.6 percent). Stated differently, while our data set is split almost evenly between pro-environmental (52.7 percent) and anti-environmental (47.3 percent) campaign promises, 66.6 percent of the defections in our data set were from pro-environmental campaign promises. These figures are roughly similar in the U.S. Senate, though overall defection rates were higher. By a 2:1 margin, then, defections moved legislator policy choice in a more anti-environmental direction.

Second, the propensity to defect varies across issue areas. We calculate defection rates for seven issue areas in environmental policy, and Table 1 shows that the probability of defection varies substantially across these issue areas. While patterns of defection across issues differ for the House and Senate, defection rates are generally higher for legislation addressing energy policy and endangered species protection, while defection rates are generally lower for legislation seeking to establish or redefine property rights or reform environmental regulation.

Third, the propensity to defect varies over time. We see that in the House the defection rate was significantly higher during the 107th Congress, while in the Senate defection peaked during the 106th Congress. In neither chamber do we see any trend in defection rates over time, and we note that the higher overall defection rate in the Senate is not a consistent phenomenon, but an artifact of the 106th and 108th Congresses. We consider two possible explanations for these differences across Congresses; though given the small sample size (6 Congresses), our results should be seen as speculative rather than definitive. First, research on parliamentary systems has found lower congruence between citizen preferences and policy choices in coalition governments than in majority governments (Blais & Bodet, 2006; Golder & Stramski, 2010). A similar phenomenon may operate in the United States, with higher defection rates during periods of divided party control over the legislature. During our time frame, legislative control was divided only during the 107th Congress, but this Congress saw the highest levels of defection in the House and the second highest in the Senate. Second, defection rates may vary as a function of the different mix of issues addressed by each Congress. To assess this explanation, for each Congress we calculated the proportion of roll call votes cast in issue areas identified in Table 1 where defection rates exceeded .40, and correlated this with the defection rate. This correlation is a robust .46, indicating that the defection rates are highest in Congresses with large proportions of bills addressing energy policy, climate change, and endangered species. This suggests that defection rates differ across Congresses not because individual propensities to defect differ over time, but because the relative frequency of roll call votes in high defection issue areas does.

Predicting Defection from Campaign Promises

  1. Top of page
  2. Abstract
  3. Introduction
  4. Campaign Promises, Representation, and Environmental Policy
  5. Connecting Campaign Promises to Policy Choices
  6. How Often Do Legislators Defect from Campaign Promises?
  7. Predicting Defection from Campaign Promises
  8. Results and Discussion
  9. Campaign Promises, SMDs, and Environmental Policy
  10. Concluding Remarks
  11. References
  12. Appendix

The Dependent Variable

The dependent variable in the prediction models is a dichotomous indicator of whether a particular legislator defected from her campaign promise (coded 1). This is the same variable used to construct Table 1. Not all members of Congress responded to the NPAT, however, and for these members we are missing information regarding defection (i.e., the dependent variable is partially unobserved). We might expect that the decision to respond to the NPAT is correlated with the decision to defect from the NPAT response—e.g., legislators intending to break a campaign promise may be unwilling to advertise the campaign promise in the first place. In this situation, NPAT responses and defection are jointly determined, and the partial observability of NPAT responses is compounded by their endogeneity. The traditional approach for dealing with partially observed, binary endogenous variables was developed by Heckman (1979; see also Puhani, 2000). Our original models of defection employed one equation predicting the propensity of a winning candidate to respond to the NPAT and a second equation predicting the propensity to defect (i.e., we employed bivariate probit models with selection). Diagnostics from these models showed that NPAT responses and defection are independent after controlling for the other variables in the defection model (see the Wald tests in Table 2). Therefore, we report the results from simple probit models in Table 2.

Table 2. Probit Coefficients for Defection from Environmental Campaign Promises, 103rd–108th Congress
Independent VariablesHouse Model 1House Model 2Senate Model 1Senate Model 2
  1. Note: Numbers in parentheses are clustered standard errors.

  2. *p < 0.05, **p < 0.01, ***p < 0.001, one-tailed tests.

  3. LCV, League of Conservation Voters.

Base model    
Pro-environmental promise−0.673***−0.392***−1.063***−0.938***
(0.039)(0.110)(0.159)(0.205)
Republican party−1.184***−0.743***−0.941***−0.929***
(0.040)(0.159)(0.159)(0.171)
Pro-environmental promise ×2.044***1.523***1.924***1.821***
Republican party(0.051)(0.142)(0.183)(0.268)
Factors enhancing representation    
Distance from constituents 0.449*** 0.239*
(0.058)(0.129)
Contributions anti-promise −0.598** 0.277
(0.238)(0.175)
Time since campaign promise  0.026
(0.023)
Factors compromising representation    
Contributions pro-promise 0.094 0.111
(0.174)(0.105)
Terms in office −0.0002 −0.007
(0.0004)(0.020)
Control variables    
Closeness of vote −0.002*** −0.005*
(0.0002)(0.002)
Majority party member −0.064 −0.122
(0.131)(0.108)
Majority party member × 0.001*** −0.001
Closeness of vote(0.0003)(0.003)
Party deviation 0.066 −0.045
(0.192)(0.449)
LCV key vote −0.011 −0.026
(0.030)(0.079)
Environmental extremism −0.006*** 0.005
(0.001)(0.004)
Constant0.105**−0.213*0.331**−0.008
(0.034)(0.109)(0.142)(0.256)
Psuedo-R20.120.150.100.12
Wald test for independence2.150.850.770.81
N14,14314,1311,3211,321

A Baseline Model for Defection from Campaign Promises

As we have seen, there is little research investigating whether elected officials make policy choices consistent with their preelection pronouncements. Consequently, there is no established body of scholarship identifying the most important factors explaining legislator defection from campaign promises. One place to begin this investigation is by considering factors that lead to representation distortion and bias more generally. For example, McDonald et al. (2004) find that distortion and bias are larger in legislatures selected through SMDs than through PR. This important discovery gives us little leverage, however, as all members of the U.S. Congress are chosen through SMDs. More useful is the research by Kim, Powell, and Fording (2010, p. 1) demonstrating that “party system polarization seems to be the predominate factor shaping distortion of governments' relationship with the median voter.” While Kim et al. were discussing the gap between citizen preferences and legislative policy positions, we might also expect that partisan polarization creates distortion between citizen preferences and legislative policy choices. Specifically, partisanship may affect the propensity to defect from campaign promises, thereby contributing to distortion between the policy preferences of voters and the policy choices of legislatures.

Party polarization has increased in the United States during the past generation. This polarization is especially noteworthy in the previously bipartisan area of environmental policy, with Democrats and Republicans staking out increasingly pro- and anti-environmental policy positions, respectively (Nelson, 2002; Shipan & Lowry, 2001). The story linking defection and party polarization is not simple and straightforward. We do not expect that members of one party will be more likely to defect from their campaign promises. Rather, increased partisanship means that members of each party will be more likely to defect from a particular type of campaign promise. Specifically, Republicans will be more likely to defect from pro-environmental campaign promises, while Democrats will be more likely to defect from anti-environmental campaign promises. Defection then serves to exacerbate preexisting partisan tendencies in environmental policy, pushing policy choices toward extreme positions. This contributes to short-term distortions between citizen preferences and policy choices of the sort that Kim et al. (2010) find with respect to legislative policy positions.

While we do not have an expectation regarding the independent effect of political party on the probability of defection, this is not true for the content of the campaign promises. Previously we offered an argument as to why U.S. legislators should be more likely to defect from pro-environmental campaign promises, and Table 1 presents evidence consistent with this argument. In our prediction model, we expect greater levels of defection from pro-environmental campaign promises.

Our baseline model predicts defection from campaign promises using partisanship and the content of the campaign promise. To test the baseline model, we use a dichotomous variable identifying Republican Representatives and Senators (coded 1), a dichotomous variable identifying pro-environmental campaign promises (coded 1), and an interaction term that is the product of these two variables. The results are presented in columns 1 and 3 of Table 2, and these results strongly support the intuition behind the baseline model. The positive coefficients on the interaction terms indicate that Republicans are more likely to defect from pro-environmental campaign promises, while the negative coefficients on the Republican Party variable indicate that these same decision makers are significantly less likely to defect from anti-environmental campaign promises. By contrast, Democrats are significantly less likely to defect from pro-environmental campaign promises (an effect captured by the pro-environmental promise coefficients) and significantly more likely to defect from anti-environmental campaign promises (an effect captured by the constant term).

The conditional effect of partisanship on the probability of defection is represented graphically in Figure 1a (for the House) and 1b (for the Senate). Figure 1a shows that for Republican members of the House of Representatives, the probability of defection rises from 14 percent for anti-environmental promises to 62 percent for pro-environmental campaign promises. For Democratic representatives, the comparable figures are 54 percent and 28 percent. Taken together, all members of the House of Representatives are more likely to defect from pro-environmental campaign promises (40 percent) than they are from anti-environmental campaign promises (22 percent). We observe a similar pattern in Figure 1b for the Senate, where Republicans are more likely to defect from pro-environmental campaign promises (60 percent to 27 percent) and Democrats are less likely to defect from pro-environmental campaign promises (23 percent to 63 percent). Overall, Senators are more likely to defect from pro-environmental campaign promises (44 percent to 34 percent).

figure

Figure 1. (a) U.S. House and (b) Senate Defections from Campaign Promises by Political Party and Valence of Promise.

Download figure to PowerPoint

A More Complete Model for Defection from Campaign Promises

Partisanship and content are not the only factors contributing to the propensity to defect from campaign promises. In an effort to specify this process more completely, we add to the base model factors that might enhance representation, factors that might compromise representation, and a set of control variables.

Factors Potentially Enhancing Effective Representation

Legislator defection from campaign promises need not always threaten popular sovereignty. We include three variables that may improve representation by influencing the propensity of defection from campaign promises. First, the policy choices of legislators are often responsive to the wishes of constituencies (Erikson & Wright, 2001; Fenno, 1978; Miller & Stokes, 1963). Therefore, we expect that legislators will be more likely to defect if their policy preferences are at odds with those of their constituents. Accurately measuring this concept, which we term electoral distance, is somewhat complicated. We measure constituency preferences using the most recent Republican candidate share of the two-party vote for president in each legislative district (designated PRep), and measure electoral distance using the difference between constituency preferences (i.e., PRep) and the Republican legislative candidate's share of the two-party vote in the most recent election (designated CRep). The effect of a legislator's distance from her constituency, however, is conditional upon the direction of the distance and the type of campaign promise that is made. Consider two examples: Legislator A is a Republican who gives a pro-environmental NPAT response and wins 51 percent of the vote in a district where the Republican presidential candidate received only 41 percent of the vote; legislator B is a Democrat who gives a pro-environmental NPAT response and wins 53 percent of the vote in a district where the Republican presidential candidate received 51 percent of the vote. Looking only at election returns, legislator A is both more vulnerable (i.e., has a smaller winning margin) and is further away from her constituency (i.e., a larger distance between her vote share and the vote share of her party's Presidential candidate) than is legislator B. We hypothesize, however, that legislator A will receive less pressure to defect because her pro-environmental campaign promise is more consistent with her constituency's preferences (where a majority voted for the Democratic presidential candidate) than is legislator B's pro-environmental response (where a majority voted for the Republican presidential candidate). To accurately capture this dynamic, our measure of electoral distance is PRep–CRep for candidates that give pro-environmental NPAT responses, and CRep–PRep for candidates that give anti-environmental NPAT responses. In either case, higher values are associated with increased pressure to defect.

Second, in U.S. elections, campaign contributions are a vital mechanism for stakeholders to communicate their preferences to candidates. If campaign contributions encourage legislators to remain faithful to their campaign promises, these contributions serve to enhance democratic representation. We only have data for campaign contributions from opponents of environmental regulation, so this representation-enhancing function of contributions can only be observed for anti-environmental campaign promises. Therefore, this independent variable takes on a value of 0 for anti-environmental campaign contributions going to candidates making pro-environmental campaign promises, and takes on the value of campaign contributions received for candidates making anti-environmental campaign promises. A negative coefficient estimate for this variable indicates that campaign promises encourage members of Congress to remain faithful to anti-environmental campaign promises.

Third, legislative candidates are asked to respond to the NPAT during each election. For members of the House, this means that there is no more than a two-year lag between their articulation of policy preferences and their policy choices, and this lag is identical for all Representatives. Given the rotating election cycle in the Senate, however, Senators may have made their relevant campaign promises anywhere from one to six years prior to the roll call vote. Objective conditions may have changed between the time the campaign promise was made and the time the roll call vote was taken, or significant policy learning may have occurred during this period. Either of these circumstances may render the original campaign promise less wise. Therefore, we expect that in the Senate the probability of defection will increase with the time elapsed since the campaign promise was made. We operationalize this expectation with a variable measuring the number of years between the NPAT response and the roll call vote.

Factors Potentially Compromising Effective Representation

We include two variables that may degrade effective representation by influencing the propensity to defect from campaign promises. First, while campaign contributions may improve the quality of representation by encouraging legislator fidelity to campaign promises, they may also compromise effective representation if they encourage legislators to defect from campaign promises. Indeed, this latter effect is more commonly stated. To test for this effect, we include a variable that takes on a value of 0 for legislators making anti-environmental campaign promises, and takes on the value of campaign contributions received from anti-environmental interests by candidates making pro-environmental campaign promises. This variable is the complement to the campaign contribution variable described earlier. The key difference is that by construction, we expect that this variable will have a positive relationship with the propensity to defect from pro-environmental campaign promises.

Second, one of the most forcefully articulated arguments for term limits is that long-time legislators “go their own way” and lose touch with their constituency. Parker (1992) formalizes this argument by claiming that over time, legislators seek to maximize their policymaking discretion. This argument indicates that long-term legislators may be more likely to defect from their campaign promises. We operationalize this expectation using the number of years each legislator has served in office.

Control Variables

We include six control variables in our full model predicting defection from campaign promises in environmental policy. First, we expect that legislators will be less likely to defect from their campaign promises when these defections might significantly affect legislative outcomes. That is, that the closeness of the roll call vote will have a negative effect on defection, all other things equal. We measure the closeness of the roll call vote by subtracting the number of votes on the winning side from the number of votes on the losing side of all roll calls, giving us a variable where larger values represent closer votes.

Second, in virtually all legislatures majority parties have more resources at their disposal to entice reluctant legislators to abandon their personal position on an issue in favor of the official party position (e.g., Cox & McCubbins, 1993). Therefore, our second control variable takes on a value of 1 for members of the majority party.

Third, the resources of the majority party are not without limit, and thus it would be irrational for the majority leadership to encourage defection on all roll call votes. Rather, we expect that the leadership will target their resources to pressure for defections on roll call votes where defections will have the greatest policy consequences: exceptionally close votes. Consequently, our third control variable interacts majority party membership with the continuous variable measuring the closeness of the roll call vote. We expect that this interaction term will generate a positive coefficient, indicating that members of the majority party are more likely to defect from their campaign promises on close votes.

Fourth, we expect that if legislators defect from their campaign promises, they will defect toward the position of their political party, and these defections will be more likely among party deviants (i.e., legislators making campaign promises at odds with the modal position of party members). We measure the position deviation of legislators in the following way. First, we calculate the proportion of legislators from each party who offer the modal (i.e., majority) response on each NPAT question, and subtract this proportion from 1. The remainder, dubbed the party position index, ranges from 0 to .5, with values closer to zero indicating greater party unity on this issue. Next, we identify legislators whose NPAT responses differ from the modal party response. We subtract the party position index from the individual candidate's position, with the absolute remainder representing individual position deviations (this score will range from 0 to 1, with larger numbers indicating greater deviation). An example helps illustrate this strategy. Assume that 90 percent of Republican legislators support limiting the designation of critical habitat under the Endangered Species Act. The party position index here would equal .10. Republican legislators who do not support this restriction are coded as 1 (the pro-environmental position), and the resulting individual position deviation would be .90. All legislators with positions identical to the majority position of their party receive deviation scores of 0.

Fifth, each year the LCV identifies a small number of key roll call votes on issues of critical importance to the environmental policy community. We expect that legislators will be less likely to defect from their campaign promises on issues of greatest salience to members of the environmental policy subsystem (i.e., on LCV key votes). We operationalize this expectation using a dichotomous variable coded 1 for LCV key votes.

Finally, our model of defection would be incomplete without considering the intensity with which legislators hold their policy preferences. We measure the intensity of preference regarding environmental policy using a folded index of LCV scores. Legislators with LCV scores at the chamber mean receive values of zero on this measure of “environmental extremism,” while legislators with very high or very low LCV scores receive values near the maximum. Presumably, candidates holding extreme positions will be less likely to defect from campaign promises in this area.

Results and Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Campaign Promises, Representation, and Environmental Policy
  5. Connecting Campaign Promises to Policy Choices
  6. How Often Do Legislators Defect from Campaign Promises?
  7. Predicting Defection from Campaign Promises
  8. Results and Discussion
  9. Campaign Promises, SMDs, and Environmental Policy
  10. Concluding Remarks
  11. References
  12. Appendix

Columns 2 and 4 in Table 2 show the results from the full models of defection in the House and Senate, respectively. We note first that the effects of partisanship and the content of the campaign promise from columns 1 and 3 remain strong and statistically significant, if somewhat attenuated, in the fully specified models of defection. Second, each of the statistically significant coefficients in columns 2 and 4 is consistent with a priori expectations.

Specifically, the probability of defection increases with the distance between a candidate's preelection policy position and the preferences of her constituency. When defecting from campaign promises, both Representatives and Senators defect toward general constituent policy references. In addition, as campaign contributions from sources hostile to environmental regulation increase, the probability that a legislator will defect from an anti-environmental campaign promise goes down. More precisely, a $100,000 increase in campaign contributions from these groups is associated with a 2 percent reduction in the probability of defecting from an anti-environmental campaign promise. The mean contribution from these groups is only $27,814, which suggests that for the typical legislator, campaign contributions have a negligible effect on the probability of defection. The maximum value of this variable is over three million dollars, however, with several legislators receiving more than one million dollars in contributions from these groups. For these legislators, our model suggests that campaign contributions are associated with a meaningful reduction in the probability of defecting from anti-environmental campaign promises. Each of the coefficients discussed in this paragraph serve to enhance the quality of representation by moving legislative policy decisions closer to constituent preferences and/or reducing the probability of defection.

Three of the control variables also display statistically significant coefficients in the expected direction. First, both Representatives and Senators are less likely to defect from campaign promises on close votes. Defection is less common when it is more consequential. The effect of the closeness of the vote is attenuated for members of the majority party in the House of Representatives, as illustrated by the negative parameter estimate on the interaction term. This result is consistent with a story that the majority party marshals its limited resources to encourage defection among members only when it is likely to change the outcome of a roll call vote. Still, the influence of majority party status is not enough to reverse the effect of vote closeness on defection, as the parameter on the interaction term is only one half the size of the vote closeness parameter. Members of the majority party are still less likely to defect as roll call votes get close; they are just more likely to defect under these circumstances than are members of the minority party, although the effect is not large. On tie votes, the probability of defection is .23 for members of the majority party and .21 for members of the minority. Finally, representatives holding intense preferences are less likely to defect from their campaign promises. Holding all other variables at their mean values, a member of the House of Representatives with a mean LCV score has a .29 probability of defection, while an identical representative with the most extreme LCV score (0 or 100) has a probability of defection equal to .24.

Some of the parameters showing no effect on the probability of defection are also of interest. For example, campaign contributions from opponents of environmental regulation do not increase the probability of defection from pro-environmental campaign promises, and long-term members of Congress are no more likely to defect from campaign promises. Observers of American politics may find these results comforting. On the other hand, fidelity to campaign promises is no more likely on especially important pieces of environmental legislation.

Campaign Promises, SMDs, and Environmental Policy

  1. Top of page
  2. Abstract
  3. Introduction
  4. Campaign Promises, Representation, and Environmental Policy
  5. Connecting Campaign Promises to Policy Choices
  6. How Often Do Legislators Defect from Campaign Promises?
  7. Predicting Defection from Campaign Promises
  8. Results and Discussion
  9. Campaign Promises, SMDs, and Environmental Policy
  10. Concluding Remarks
  11. References
  12. Appendix

Social scientists have long recognized that the policy choices of democracies differ from those of autocracies in areas as varied as trade (Mansfield, Milner, & Rosendorf, 2000), currency exchange rates (Freeman, Hays, & Stix, 2000), and the use of force (Maoz & Russett, 1993). Congleton (1992) brought this perspective to the study of environmental policy, concluding that democracies adopt more stringent domestic environmental policies than do autocracies (see also Bättig & Bernauer, 2009; Neumayer, 2002; Roberts, Parks, & Vasquez, 2004). The effect of democracy on public policy is not constant, however, and the design of democratic institutions has a meaningful effect on the policy choices of governments (e.g., Congleton & Swedenborg, 2006; Lijphart, 1999). Of particular relevance for us is the observation that legislatures selected through SMDs adopt less stringent environmental policies than do legislatures selected through PR. For example, when examining the electoral rules governing national legislatures in 86 democracies, Fredriksson and Millimet (2004) conclude that SMD legislatures adopt lower fuel taxes, participate in fewer international environmental agreements, and invest less in environmental governing capacity. Lijphart (1999) and Poloni-Staudinger (2008) also note that majoritarian systems characterized in part by SMD legislatures adopt weaker environmental policies, and Fredriksson and Wollscheid (2007) go so far as to conclude that the positive relationship between democracy and the stringency of domestic environmental policy is driven almost exclusively by Parliamentary democracies with PR legislatures.

The observation that SMD systems like the United States adopt weaker environmental policies begs the question of why this might be so. Persson and Tabellini (1999) offer one explanation. Political parties in SMD systems can succeed by appealing to the preferences of a subset of their constituency. At the extreme, one party can fully populate the legislature by representing the interests of barely 50 percent of voters. By contrast, party success in PR systems is maximized by appealing to as broad a constituency as possible (see also Milesi-Ferretti, Perotti, & Rostagno, 2002).6 One implication from this work is that legislators in SMD systems will emphasize policies providing local public goods and appealing to particularistic interests, while PR systems provide greater incentives for legislators to pay attention to aggregate social welfare and global public goods (see also Cox & McCubbins, 2001). As environmental policy typically generates concentrated local costs and diffuse national benefits, legislators in SMD systems have a weaker incentive to advocate for and adopt stringent environmental policies (Fredriksson & Wollscheid, 2007).

We believe that legislator defection from campaign promises may offer a second, complementary explanation for why SMD systems adopt weaker environmental policies. Two conditions must be satisfied if defection from campaign promises is to play this complementary role. First, defection from campaign promises must move policy decisions toward weaker environmental policies. The analysis here provides a rationale for why defection should be more common for pro-environmental campaign promises, and provides evidence that defection is more common for pro-environmental campaign promises. For the United States, at least, this first condition appears to be satisfied.

Second, defection must be more common in SMD systems than in PR systems. Offering a plausible rationale for why defection is more common in SMD systems is not difficult. In PR systems, the notion of campaign promises made by individual candidates is something of a non sequitur, as legislators in PR systems tend to be selected from party lists. Moreover, the strong party discipline traditionally exercised in PR systems means that defection from party policy positions is unusual. Coupled with strong party discipline, the corporatist character of many PR systems means that legislators face less pressure from interest groups seeking to influence their policy choices. By contrast, SMD systems typically have candidate-centered campaigns, weaker party discipline, and pluralist interest representation. Each of these characteristics increases the probability that legislators in SMD systems will make and defect from campaign promises, compared with their counterparts in PR systems. While there is little empirical evidence regarding the extent to which legislators defect from their campaign promises, the evidence we do have suggests that defection is indeed more likely in SMD systems. For example, studying the U.S. House of Representatives in the 105th Congress, Ringquist and Dasse (2004) find that legislators defected from their campaign promises 27 percent of the time.7 Moreover, Table 2 shows that defection from environmental campaign promises in the United States ranges from 31 percent in the House to 41 percent in the Senate. By contrast, a comparable analysis of the Swiss legislature—where a large majority of representatives are chosen via PR—found defection rates of only 15 percent (Schwarz, Schädel, & Ladner, 2010). More convincingly, a recent summary of the small empirical literature on this question concludes that compared with presidential systems like that in the United States, “parliamentary regimes like the Westminster systems of Britain and Canada positively influence the likelihood that political parties keep their promises once elected” (Pétry & Collette, 2009, p. 78). Therefore, the second condition may be satisfied as well.

We do not want to overstate the case—the results in Table 2 do not provide evidence that defection from campaign promises accounts for why SMD systems adopt weaker environmental policies. We are simply pointing out that if the high levels of defection from pro-environmental campaign promises in the United States also characterize other SMD systems, this might help account for this phenomenon.

Concluding Remarks

  1. Top of page
  2. Abstract
  3. Introduction
  4. Campaign Promises, Representation, and Environmental Policy
  5. Connecting Campaign Promises to Policy Choices
  6. How Often Do Legislators Defect from Campaign Promises?
  7. Predicting Defection from Campaign Promises
  8. Results and Discussion
  9. Campaign Promises, SMDs, and Environmental Policy
  10. Concluding Remarks
  11. References
  12. Appendix

We began this manuscript with two goals: measuring the extent to which candidates for the U.S. Congress defect from their campaign promises in environmental policy, and investigating whether these defections are systematic. Regarding the first goal, when given the opportunity to fulfill their campaign promises in environmental protection, members of the House of Representatives defect from these promises 31 percent of the time. The figure for Senators is 41 percent. While legislators in both chambers exhibit fidelity to preelection policy preferences in a clear majority of cases, defection is a common characteristic of legislative behavior in the United States. Regarding the second goal, defection from campaign promises is a systematic phenomenon. The propensity for defection is dominated by partisanship and the content of the campaign promise. Overall, defection from pro-environmental campaign promises is twice as common as defection from anti-environmental campaign promises, and Republicans are more likely to defect from pro-environmental campaign promises while Democrats are more likely to defect from anti-environmental pronouncements. In addition, legislators defect toward the general ideological position of their constituents, and are marginally less likely to defect when they hold extreme policy positions or when the consequences of defection for policy choice are greatest.

Our results have implications for the quality of democratic governance in environmental policy. First and most directly, defection from campaign promises clearly undercuts effective democratic representation and poses a challenge to the applicability of mandate theory as a mechanism for linking citizen preferences to environmental policy choices in the United States. Defection from campaign promises also shapes the direction of environmental policy. Because legislators in the United States are significantly more likely to defect from pro-environmental campaign promises, defection has the effect of moving government policy toward weaker environmental programs. Thus, defection from campaign promises may help account for the observation that the United States adopts weaker environmental protection policies than its citizens might prefer. Second, defection from campaign promises may be one mechanism behind the observation that SMD systems adopt less stringent and less comprehensive environmental policies. The ultimate contribution of defection to observed differences in environmental policy between SMD and PR democracies hinges on the extent to which defection rates in the United States are representative of legislative defection in other SMD systems.8 We leave the answer to this question to future research.

Notes
  1. 1

    This perspective closely parallels what Mansbridge (2003) calls “promissory” representation in which candidates seek to attract voters by making promises regarding future policy choices, and voters evaluate candidates based upon those promises.

  2. 2

    The 1992 NPAT asked candidates to take positions on three environmental policy proposals. All other NPATs asked far more environmental policy questions (9 questions in 1994, 12 questions in 1996, 12 questions in 1998, 11 questions in 2000, and 10 questions in 2002).

  3. 3

    In 2008 the NPAT was renamed the “Political Courage Test.”

  4. 4

    Despite this evidence, some readers may still prefer to interpret NPAT responses as “public statements of policy preferences coupled with a promise of action,” rather than as campaign promises per se. This interpretation has no effect on the analysis or conclusions in the article.

  5. 5

    Defining defection by matching legislator votes to campaign promises implicitly assumes that legislators vote sincerely. Despite the widespread expectation in the public choice literature that legislators will instead vote strategically, the empirical evidence indicates that strategic voting is extraordinarily rare. Groseclose and Milyo (2010) find that by making a minor plausible change to the assumptions of public choice models, the models predict that legislators will almost always vote sincerely rather than strategically.

  6. 6

    This is not the only strategy for legislative success in PR systems. For example, narrowly focused small parties may be more successful at exceeding thresholds for representation than are more generalist small parties. The argument is simply that in PR systems, there is a stronger positive relationship between constituency support and the share of seats controlled in the legislature.

  7. 7

    Sullivan and O'Connor (1972) and Fishel (1985) also examine whether U.S. elected officials keep their campaign promises, though their work is not as applicable to the question asked here. Fishel examines Presidential promise keeping, and Sullivan and O'Connor examine the correlation between policy statements and roll call votes, but do not measure the extent of defection from campaign promises.

  8. 8

    The propensity to make and break campaign promises may be correlated with the incentives of legislators to cultivate a “personal vote” (Carey & Shugart, 1995; Crisp, Escobar-Lemmon, Jones, Jones, & Taylor-Robinson, 2004). Incentives for cultivating personal votes are high in the United States, and on a par with Taiwan, Columbia, the Philippines, Brazil, and Finland. While our results may be most generalizable to these countries, ultimately this is an empirical question.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Campaign Promises, Representation, and Environmental Policy
  5. Connecting Campaign Promises to Policy Choices
  6. How Often Do Legislators Defect from Campaign Promises?
  7. Predicting Defection from Campaign Promises
  8. Results and Discussion
  9. Campaign Promises, SMDs, and Environmental Policy
  10. Concluding Remarks
  11. References
  12. Appendix
  • Aldrich, John H., John L. Sullivan, and Eugene Borgida. 1989. “Foreign Affairs and Issue Voting: Do Presidential Candidates ‘Waltz Before a Blind Audience'?American Political Science Review 83 (1): 123141.
  • Ansolabehere, Stephen, and Phillip Edward Jones. 2010. “Constituents Response to Congressional Roll Call Voting.” American Journal of Political Science 54 (3): 583597.
  • Ansolabehere, Stephen, James M. Snyder Jr., and Charles Stewart III. 2001a. “Candidate Positioning in the U.S. House Elections.” American Journal of Political Science 45 (1): 136159.
  • Ansolabehere, Stephen, James M. Snyder Jr., and Charles Stewart III. 2001b. “The Effects of Party and Preferences on Congressional Roll-All Voting.” Legislative Studies Quarterly 26 (4): 533572.
  • Barrett, Edith J., and Fay Lomax Cook. 1991. “Congressional Attitudes and Voting Behavior: An Examination of Support for Social Welfare.” Legislative Studies Quarterly 16 (3): 375392.
  • Bättig, Michèle B., and Thomas Bernauer. 2009. “National Institutions and Global Public Goods: Are Democracies More Cooperative in Climate Change Policy?International Organization 63 (2): 281308.
  • Blais, André, and Marc André Bodet. 2006. “Does Proportional Representation Foster Closer Congruence Between Citizens and Policy Makers?Comparative Political Studies 39 (10): 12431262.
  • Buttice, Mathew, and Walter Stone. 2012. “Candidates Matter: Policy and Quality Differences in Congressional Elections.” Journal of Politics 74 (3): 870887.
  • Carey, John M., and Matthew Soberg Shugart. 1995. “Incentives to Cultivate a Personal Vote: A Rank Ordering of Electoral Formulas.” Electoral Studies 14 (4): 417439.
  • Congleton, Roger D. 1992. “Political Institutions and Pollution Control.” Review of Economics and Statistics 74 (3): 412421.
  • Congleton, Roger D. , and Birgitta Swedenborg , eds. 2006. Democratic Constitutional Design and Public Policy. Cambridge, MA: MIT Press.
  • Cox, Gary W., and Mathew D. McCubbins. 1993. Legislative Leviathan: Party Government in the House. Berkeley, CA: University of California Press.
  • Cox, Gary W., and Mathew D. McCubbins. 2001. “The Institutional Determinants of Economic Outcomes.” In Presidents, Parliaments, and Policy, ed. Gary Cox , and Stephen Haggard . New York: Cambridge University Press, 2163.
  • Crisp, Brian F., Maria C. Escobar-Lemmon, Bradford S. Jones, Mark P. Jones, and Michelle M. Taylor-Robinson. 2004. “Vote-Seeking Incentives and Legislative Representation in Six Presidential Democracies.” Journal of Politics 66 (3): 823846.
  • Dalton, Russell J. 2005. “The Greening of the Globe? Cross-National Levels of Environmental Group Membership.” Environmental Politics 14 (4): 441459.
  • Disch, Lisa J. 2011. “Toward a Mobilization Conception of Democratic Theory.” American Political Science Review 105 (1): 100114.
  • Dunlap, Riley E. 1995. “Public Opinion and Environmental Policy.” In Environmental Politics and Policy, 2nd ed., ed. James P. Lester . Durham, NC: Duke University Press, 63114.
  • Dunlap, Riley E., George H. Gallup Jr., and Alec M. Gallup. 1999. “Of Global Concern: Results of the Health of the Planet Survey.” In Handbook of Global Environmental Policy and Administration, ed. Dennis S. Soden , and Brent R. Steel . New York: Marcel Dekker, 728.
  • Erikson, Robert S., and Gerald C. Wright Jr. 2001. “Voters, Candidates and Issues in Congressional Elections.” In Congress Reconsidered, ed. Lawrence C. Dodd , and Bruce I. Oppenheimer . Washington, DC: Congressional Quarterly Press, 6795.
  • Fenno, Richard F. 1978. Home Style: House Members in Their Districts. Boston: Little, Brown.
  • Fishel, Jeff. 1985. Presidents & Promises: From Campaign Pledge to Presidential Performance. Washington, DC: Congressional Quarterly Press.
  • Fredriksson, Per G., and Daniel L. Millimet. 2004. “Electoral Rules and Environmental Policy.” Economics Letters 84 (2): 237244.
  • Fredriksson, Per G., and Jim Wollscheid. 2007. “Democratic Institutions vs. Autocratic Regimes: The Case of Environmental Policy.” Public Choice 130 (3–4): 381393.
  • Freeman, John R., Jude C. Hays, and Helmut Stix. 2000. “Democracy and Markets: The Case of Exchange Rates.” American Journal of Political Science 44 (3): 446463.
  • Golder, Matt, and Jacek Stramski. 2010. “Ideological Congruence and Electoral Institutions.” American Journal of Political Science 54 (1): 90106.
  • Groseclose, Timothy, and Jeffrey Milyo. 2010. “Sincere Versus Sophisticated Voting in Congress: Theory and Evidence.” Journal of Politics 72 (1): 6073.
  • Hall, Richard L. 1996. Participation in Congress. New Haven, CT: Yale University Press.
  • Heckman, James J. 1979. “Sample Selection Bias as Specification Error.” Econometrica 47 (1): 153161.
  • Hill, Kim Quaile. 2001. “Multiple-Method Measurement of Legislators Ideologies.” Legislative Studies Quarterly 27 (2): 263274.
  • Jones, Phillip Edward. 2011. “Which Buck Stops Here? Accountability for Policy Positions and Policy Outcomes in Congress.” Journal of Politics 73 (3): 764782.
  • Kahn, Kim Fridkin, and Patrick Kenney. 2001. “The Importance of Issues in Campaigns: An Examination of Citizens' Reception of Issue Messages.” Legislative Studies Quarterly 26 (4): 573597.
  • Kamieniecki, Sheldon. 2006. Corporate America and Environmental Policy: How Often Does Business Get Its Way? Stanford, CA: Stanford University Press.
  • Kim, HeeMin, G. Bingham Powell Jr., and Richard C. Fording. 2010. “Electoral Systems, Party Systems, and Ideological Representation.” Comparative Politics 42 (2): 167185.
  • Klotz, Robert. 1997. “Positive Spin: Senate Campaigning on the Web.” PS: Political Science and Politics 30 (3): 482486.
  • Kraft, Michael. 2006. Environmental Policy and Politics, 4th ed. New York: Pearson Longman.
  • Lijphart, A. 1999. Patterns of Democracy. New Haven, CT: Yale University Press.
  • Lodge, Milton, and Marco Steenbergen, with Shawn Brau. 1995. “The Responsive Voter: Campaign Information and the Dynamics of Candidate Evaluation.” American Political Science Review 89 (2): 309326.
  • Mansbridge, Jane. 2003. “Rethinking Representation.” American Political Science Review 97 (4): 515528.
  • Mansfield, Edward D., Helen V. Milner, and B. Peter Rosendorf. 2000. “Free to Trade: Democracies, Autocracies, and International Trade Negotiations.” American Political Science Review 94 (2): 305322.
  • Maoz, Zeev, and Bruce Russett. 1993. “Normative and Structural Causes of Democratic Peace, 1946–1986.” American Political Science Review 87 (3): 624638.
  • McDonald, Michael D., Silvia M. Mendes, and Ian I. Budge. 2004. “What Are Elections for? Conferring the Median Mandate.” British Journal of Political Science 34 (1): 126.
  • Mertig, Angela G., and Riley E. Dunlap. 2001. “Environmentalism, New Social Movements, and the New Class: A Cross-National Investigation.” Rural Sociology 66 (1): 113136.
  • Milesi-Ferretti, Gian Maria, Roberto Perotti, and Massimo Rostagno. 2002. “Electoral Systems and Public Spending.” Quarterly Journal of Economics 117 (2): 609657.
  • Miller, Warren E., and Donald E. Stokes. 1963. “Constituency Influence in Congress.” American Political Science Review 57 (1): 4556.
  • Nelson, Jon P. 2002. “ ‘Green' Voting and Ideology: LCV Scores and Roll-Call Voting in the U.S. Senate, 1988–1998.” Review of Economics and Statistics 84 (3): 518529.
  • Neumayer, Eric. 2002. “Do Democracies Exhibit Stronger International Environmental Commitment? A Cross Country Analysis.” Journal of Peace Research 39 (2): 139164.
  • Parker, Glenn R. 1992. Institutional Change, Discretion, and the Making of Modern Congress. Ann Arbor, MI: University of Michigan Press.
  • Persson, Torsten, and Guido Tabellini. 1999. “The Size and Scope of Government: Comparative Politics with Rational Politicians.” European Economic Review 43 (4–6): 699735.
  • Pétry, François, and Benoît Collette. 2009. “Measuring How Political Parties Keep Their Promises: A Positive Perspective from Political Science.” In Do They Walk Like They Talk?, ed. Louis M. Imbeau . Studies in Public Choice 15. New York: Springer, 6580.
  • Poloni-Staudinger, Lori. 2008. “Are Consensus Democracies More Environmentally Effective?Environmental Politics 17 (3): 410430.
  • Project Vote Smart Website. 2007. [Online]. http://www.vote-smart.org. Accessed April 2, 2013.
  • Puhani, Patrick. 2000. “The Heckman Correction for Sample Selection and Its Critique.” Journal of Economic Surveys 14 (1): 5367.
  • Ringquist, Evan J., and Carl Dasse. 2004. “Lies, Damned Lies, and Campaign Promises: Environmental Legislation in the 105th Congress.” Social Science Quarterly 85 (2): 400419.
  • Roberts, J. Timmons, Bradley C. Parks, and Alexis A. Vasquez. 2004. “Who Ratifies Environmental Treaties and Why? Institutionalism, Structuralism, and Participation by 192 Nations in 22 Treaties.” Global Environmental Politics 4 (3): 2264.
  • Rosenbaum, Walter A. 2010. Environmental Politics and Policy, 8th ed. Washington, DC: Congressional Quarterly.
  • Schwarz, Daniel, and Lisa L. Schädel, and Andreas Ladner. 2010. “Pre-Election Positions and Voting Behavior in Parliament: Consistency among Swiss MPs.” Swiss Political Science Review 16 (3): 533564.
  • Shipan, Charles, and William Lowry. 2001. “Environmental Policy and Party Divergence in Congress.” Political Research Quarterly 54 (2): 245264.
  • Shor, Boris, Christopher Berry, and Nolan McCarty. 2010. “A Bridge to Somewhere: Mapping State and Congressional Ideology on a Cross-Institutional Common Space.” Legislative Studies Quarterly 35 (3): 417448.
  • Shor, Boris, and Nolan McCarty. 2011. “The Ideological Mapping of American Legislatures.” American Political Science Review 105 (3): 530551.
  • Sulkin, Tracy. 2005. Issue Politics in Congress. New York: Cambridge University Press.
  • Sullivan, John L., and Robert E. O'Connor. 1972. “Electoral Choice and Popular Control of Public Policy: The Case of the 1966 House Elections.” American Political Science Review 66 (4): 12561268.
  • Switzer, Jaqueline Vaughn. 2001. Environmental Politics: Domestic and Global Dimensions. Boston: Bedford/St. Martins.
  • Toms, Michael, and Robert Van Houweling. 2008. “Candidate Position and Voter Choice.” American Political Science Review 102 (3): 303318.
  • Vig, Norman , and Michael Faure , eds. 2004. Green Giants? Environmental Policies of the United States and the European Union. Cambridge, MA: MIT Press.
  • West, Darrell. 1993. Air Wars: Television Advertising in Election Campaigns, 1952–1992. Washington, DC: Congressional Quarterly.
  • Wright, Gerald C., and Michael B. Berkman. 1986. “Candidates and Policy in United States Senate Election.” American Political Science Review 80 (2): 567588.

Appendix

  1. Top of page
  2. Abstract
  3. Introduction
  4. Campaign Promises, Representation, and Environmental Policy
  5. Connecting Campaign Promises to Policy Choices
  6. How Often Do Legislators Defect from Campaign Promises?
  7. Predicting Defection from Campaign Promises
  8. Results and Discussion
  9. Campaign Promises, SMDs, and Environmental Policy
  10. Concluding Remarks
  11. References
  12. Appendix
Panel A: House
Variable NameMinMeanMaxSDN
Defection00.41710.49338,350
Pro-environmental promise00.52810.49914,577
Republican party00.52210.50039,585
Distance from constituents00.11410.31739,585
Anti-promise contributions027,8153,312,89286,73039,585
Pro-promise contributions020,8893,669,69583,76139,585
Terms in office15.123263.92939,585
Closeness of vote–396–88.73075.8839,150
Majority party member00.52310.49939,585
Party deviation00.0430.4750.10314,577
LCV key vote00.43310.49639,150
Environmental extremism 0.05031.3553.9513.1939,503
Panel B: Senate
Variable NameMinMeanMaxSDN
Defection00.50510.5006,041
Pro-environmental promise00.67710.4681,339
Republican party00.51610.5006,200
Distance from constituents00.10910.3116,200
Anti-promise contributions049,4163,053,459211,5196,200
Pro-promise contributions0105,4913,007,326320,5186,200
Time since campaign promise12.99861.6141,339
Terms in office12.61091.6676,200
Closeness of vote–96–22.40020.116,200
Majority party member00.52710.4996,200
Party deviation00.0350.4550.0951,339
LCV key vote00.56510.4966,200
Environmental extremism0.3634.5855.6412.956,199