Arkansas Tech University The Culture Wars & Political Polarization in Perspective
Download 4.38 Mb.
|
|
Table 9.5: Weighted & Unweighted Mean Deviation Trend Regressions of PID GP on Ideology, 1972-2004
contributors to group polarization in the 1970’s relative to the other party identifiers while less so in the other, more recent, decades. Republican relative contribution has been on the decline. In Table 9.5, I regress the weighted and unweighted group polarization score percent contributions represented as a deviation from the mean contribution for the time series for each of the identifier groups. The model uses robust standard errors. Also included in the models is the party of the presidential administration. The reason to include presidential partisanship should be apparent. In one respect, it is indirectly influenced by the number of party identifiers. Secondly, we could see a greater contribution for out-of-power parties to partisan polarization on ideology. It is certainly plausible that Democrats may polarize in response to a Republican presidency, and likewise for Republican identifiers during a Democratic administration. In the weighted models, there are no significant linear trends in the percent contribution of Republicans or Democrats. Nor is there a significant relationship between the percent contribution of the partisan identifiers and presidential partisanship. There is a strong statistical relationship between survey year and the percent contribution of Independents, reflective of the negative trend in that categories contribution to the group polarization score. As I noted previously, this is largely the consequence of the declining number of Independent identifiers since the 1970’s. For every survey year, the contribution of Independent identifiers declines by -0.111, and the model accounts for 67.4% of the variation in contribution. There is a significant, negative relationship between the presidential administration partisanship and the contribution of Independent identifiers to the group polarization score (-0.772). This is likely a consequence of the fact that a great deal of the decline in Independent identifiers occurred in the 1990’s, when President Clinton was in office. When we treat the groups equally, ignore the relative size of the groups, there are an interesting set of patterns not apparent for the weighted group polarization measures. Just looking at the relative distances between the categories, there is a significant positive trend for the Democratic party and its contribution to polarization over the time series. For every survey year there is a 0.221 increase in the percent contribution of Democrats above the mean contribution. Democrats have become increasingly distant from the other identifiers on ideology, contributing more to group polarization. Interestingly, there is a significant negative trend for Republican identifiers (-0.164) along with the still-negative trend for Independents (-0.057). Republicans are contributing less to polarization over the full time series. The presidential partisanship effect on group polarization is in the expected direction and significant in all models. For the Republican percent contribution to group polarization, presidential partisanship is negative, declining -2.318 as the percent contribution for Republicans increases. In other words, Republicans are contributing more to group polarization during Democratic administrations. Similarly, the coefficient for presidential partisanship is positive for Democratic identifiers. Democrats are contributing more to group polarization during Republican administrations. It seems apparent that out-of-power parties react to the other party controlling the White House by shifting towards the ideological extremes. These models account for between 25% and 31% of the variance in percent contribution as measured as a deviation from the mean contribution to unweighted group polarization.
In 2004, President Bush won a narrow but decisive victory over the Democratic candidate, John Kerry. Republicans not only secured the White House, but also picked up seats in the House of Representatives (3) and in the United States Senate (4). President Bush felt emboldened to pursue a significant political agenda in the wake of this election. In a press conference shortly after the election he announced: “Let me put it to you this way: I earned capital in the campaign, political capital, and now I intend to spend it….I earned some capital. I’ve earned capital in this election—and I’m going to spend it for what I told the people I’d spend it on, which is—you’ve heard the agenda: Social Security, and tax reform, moving this economy forward, education, fighting and winning the war on terror.” Six months later, the president’s plan to partially privatize social security through personal accounts for younger workers lay in shambles(Ortiz 2007). Despite a unified Republican government, Bush’s primary public policy agenda item had failed, barely into his second term. Why? And was this unique to just Bush, or did this failure have broader implications? This wasn’t the first time an elected president, with a unified Congress, failed to get even incremental action on the primary policy goal for his administration. Bill Clinton, elected in 1992 to the presidency and with a 56 to 44 seat majority in the Senate and while losing a few seats (9), maintained a cavernous 91 seat margin over the Republicans (258-176). Again, the newly elected candidate had campaigned hard on an issue (in this case, health care), and sought to enact significant reform on the issue in the wake of his electoral victory. The plan did not disturb the employer-funded health insurance the current system is built on, however it created an overarching governmental health care superstructure (health maintenance organizations) which would mandate universal health care and regulate health care coverage. Despite unified government, by September 1994, Senate Majority Leader George Mitchell declared health care reform (having been dubbed by opponents as “HillaryCare”) dead (Bok 2003). What happened to these two presidents? Supposedly at the height of their power, with few institutional obstacles in their path and control of the entire federal government, these presidents failed to even obtain incremental change. It is not as if the health care or social security status quos are immune to change. Social Security reforms were enacted in 1977 and 1981. Medicare and Medicaid were not only instituted, but have been modified several times since the programs were initially setup. What is the answer? Some attribute it to powerful interest groups. The health care industry spent millions on opposition ads, including the famous “Harry and Louise” ads which depicted ‘regular’ citizens with worries over Clinton’s plan. The AARP was firmly aligned against President Bush’s social security plan. While these factors were certainly important, the overarching factor, I believe, was the increase in partisan polarization. The parties in the electorate had polarized, producing more ‘landslide’ districts, resulting in increasingly ideologically divergent parties in Congress. In the polarized era, advocates of significant change to the status quo face an already interested and passionate opposition that constitutes a substantial proportion of the public. Newt Gingrich, writing in the wake of President Bush’s social security plan bogging down in the legislative process, argued, “I don’t think [President Bush] can get complex reform through…It’s too hard with the AARP opposing you and all of the Democrats lined up against it.” Likewise, Hillary Clinton reported being a bit wiser in the wake of health care defeat, “I learned some valuable lessons about the legislative process, the importance of bipartisan cooperation and the wisdom of taking small steps to get a big job done.” But was there “bipartisan cooperation” to be had? Unified government was a sufficient basis for significant policy change in the past. What about unified government in the 1990’s and the first decade of the 21st century had changed from the halcyon days of unified government under Franklin Roosevelt and Lyndon Johnson—both successful in producing major policy changes and instituting significant new programs. Indeed, presidents have accomplished more legislative success under divided government, such as the major reductions in the marginal tax rates achieved by President Reagan in the 1980’s. These anecdotes suggest two aspects of partisan polarization. Intractable conflict has become more likely in competition over the policy agenda because the distance between the starting points on the two sides has become greater. Hillary Clinton may have learned the value of bipartisanship, but that doesn’t mean that elected officials are more likely to propose changes to the policy status quo likely to induce it. A polarized electorate elects polarized candidates and officials. Even if they want to compromise, the shadow of future elections where a more ideologically consistent and polarized constituency is more likely to punish them for their heresy. Witness the fate of Pennsylvania Senator Arlen Specter, who was forced to switch parties as a consequence of his vote for President Obama’s stimulus bill. Facing a strong primary challenge for his seat in Pat Toomey, Specter faced sure defeat in the 2010 Republican primary. Partisan polarization reduces the compromise space in two ways: 1) It forces the policy agenda of the party in power in the direction of their more extreme members rather than the center. 2) The opposition’s rejection zone is further distant from the center. In Figure 9.1, I depict a hypothetical scenario where bipartisan compromise on a policy change from the status quo is possible. I represent the continuum of possible policies as a line, on which the Republican and Democratic Parties have a subset of that continuum which are acceptable policies from the perspective of the parties in Congress (would command a majority of the party identifiers). I characterize the “rejection space” as the policy choices that exist outside the range of acceptable policies for either party. In the Figure 9.1 hypothetical, the status quo policy is within the range of acceptable policies for the Republicans, but outside of the range of acceptable policies for the Democrats. In this scenario, Democrats have a majority in the House, and thus can outvote the Republicans in a party-line vote. However, given not only the relative proximity of the range of potential policies but also the overlap in the two acceptable policy ranges, the possibility for a bi-partisan compromise is apparent. Especially on major and highly visible policies, the preference for bipartisan action over party-line policies should be apparent. Majority parties have an incentive to avoid stark policy choices between the two parties, as it could pave the way for electoral defeat in the next election cycle. Bringing the other party in on a policy change (in this case a policy alternative proposed within the compromise space) guards against the use of that policy by the minority party as a cudgel. In this scenario, the Democrats can propose a policy alternative that moves the status quo into the range of their acceptable policies, but also garners support from the Republicans. A bipartisan compromise and a successful change in the status quo is the result. In Figure 9.2, I construct a hypothetical that is represents a stark break from that of Figure 9.2. In this scenario, the two parties are polarized on the relevant issue dimension (or in the general ideological dimension). Note that the status quo lies outside both parties range of acceptable policies. Compromise Space
Status Quo Policy
Rejection Space Rejection Space Republican Policy Range Democratic Policy Range However, unlikely the Figure 9.1 hypothetical where a compromise was possible in the center of the dimension, the center in this case is entirely subsumed within the rejection space. Neither party finds moderate polices an acceptable alternative. There is no overlapping compromise space. Republican policy proposals will exclusively consist of polices that move the status quo substantially to the Right, while Democratic proposals will only include policies well to the Left of the center of the policy dimension. The likely result: intractable conflict (i.e. no change in the status quo) or a party-line policy imposed by the majority party. That policy will be a much more significant change in the status quo than those likely in the Figure 9.1 hypothetical. As such, it may result in no change in the status quo at all: members of the majority party may be unwilling to enact such a substantial change in the status quo given that they could be held responsible for that vote in their less-safe districts in the following election. At the same time, there is little incentive to propose moderate policies because a) the ideological extreme members of the majority party are unlikely to support it and b) it will not garner support from any of the Republicans, thus gaining bipartisan support. While it is only anecdotal evidence, there is some indication this is the kind of scenario operative in the defeat of Bill Clinton’s health care plan in 1993, the defeat of President Bush’s proposal to reform social security in 2005, and there are strong similarities between the hypothetical and the ongoing debate over President Obama’s proposal to reform health care. That all three involved unified government suggests perhaps that the presidents believed that this would allow them to enact a
Policy
Democratic Policy Range Republican Policy Range stronger change, given their control of the government. Note that all three proposals represented significant changes to the status quo. Bush sought a privatization of social security, while Obama and Clinton sought to bring health care providers, accountable for 1/7th of the economy, under the purview of government oversight. Compromise measures failed to garner the support of ideologically extreme members of the majority party. And ultimately (in the case of Bush and Clinton), there was no change to the status quo policy. There is a great deal of the debate over Obama’s health care proposal reminiscent of the Figure 9.2 hypothetical. Obama’s health proposals, including a public option for health insurance designed to ‘compete’ against private insurers, represents a significant change in the status quo. And to this point, compromise policies have a) failed to garner Republican votes and b) been resoundingly rejected by the ideological extremists in the Democratic Party. Witness the House Democrats insistence on the inclusion of the public option in their proposals despite the fact that it is considered a non-starter in the Senate. The Speaker of the House, Nancy Pelosi insists: “There is no division. We all [House Democrats] support a public option” (Miller 2009). Whether or not the result of this will be no change in the status quo and yet another failed “presidential mandate” for policy change remains to be seen. But the signs suggest that may be the path that health care reform is on. Of course, these examples constitute just anecdotal evidence suggesting that partisan polarization has reduced or eliminated the compromise space and generated intractable political conflict. A more systematic examination of policy proposals to change the status quo is beyond the scope of this analysis, but it constitutes the next step in assessing the relationship. However, there are essential predicates to the scenarios I’ve described: namely that partisan polarization has occurred at the electoral level, that this has produced partisan polarization in the elites (a reciprocal relationship would also be consistent with the scenarios), and thus it drives presidents to propose more radical Figure 9.3: Partisan Polarization in Congress, 1954-2004 
changes to the status quo in response to their polarized constituents, it produces more polarized elected officials voting on legislation, and thus leading to more conflict and less compromise. I have already shown that there is partisan polarization in the electorate. Next, I will show the partisan polarization at the elite level and assess the relationship between partisan polarization at the mass level and elite partisan polarization. Figure 9.3 reports the difference between the average D-W Nominate score for Republicans and Democrats from 1954 to 2004. The D-W Nominate score (first dimension) is generally interpretable as the Right-Left ideological dimension. Note that the difference between the two parties has substantially increased since the 1950’s in both the House of Representatives and the Senate. From an average difference 0.40 for the Senate and 0.49 for the House in 1954, the difference fifty years later has doubled, with an ideological difference between the parties in the House above 0.90 and an ideological difference in the Senate of 0.80. The Senate partisan polarization is particularly striking, given the conventional wisdom that partisan polarization in the House is exclusively due to gerrymandering and the creation of ‘landslide’ districts in recent decades. The only change along the lines of district change in the Senate was the inclusion of the states of Alaska and Hawaii to the Union, adding four new senators to the chamber. Otherwise, the territories from which senators are elected has remained fixed since 1954. Despite this, the polarization in the Senate has been nearly as substantial as that for the House. In order to compare the partisan polarization in Congress at the elite level with the observed partisan polarization at the mass level as captured by the group polarization measure, I convert the House and Senate average Nominate differences to Z-scores based on their deviation from the time-series global mean for each of the variables for the period extending from 1972 to 2004. Thus the scores for each year on partisan polarization in Congress are the z-distribution standardized scores. I convert the group polarization score to a Z-Score using the same method. With both elite and partisan polarization scored on the same scale, a direct comparison of their trends is possible (Figure 9.4). The z-score calculation is reported in equation 9.1. Figure 9.4: Trends in Z-Standardized Partisan Polarization at the Mass and Elite Levels, 1972-2004 
Equation 9.1: Z-Score for Partisan Polarization ZPP = Where:
 = the ith observed value of partisan polarization for  year.
 = the ith mean value of partisan polarization for year.
 = the ith standard deviation of partisan polarization for year.
As is apparent in Figure 9.4, there has been significant and substantial partisan polarization at both elite and mass levels since 1972. Furthermore, the magnitude of the polarization at the mass level tracks closely with polarization at the elite level. The negative Z-scores for the congressional D-W Nominate (ideological) differences and the negative Z-scores for the mass partisan polarization are in the 1970’s and the 1980’s. In 1992, the partisan polarization for both elites and masses crosses over to positive standardized partisan polarization, with the partisan polarization at the elite and mass levels in 2004 at or near two standard deviations above the mean. Recall, the negative z-scores do not indicate depolarization or moderation, but rather these partisan polarization scores fall below the global mean. Table 9.5 reports regression models using mass partisan polarization to predict elite partisan polarization. Note the strong relationship between the trend in partisan polarization for the public and the difference in ideology for the Republican and Democratic Parties in both the House and the Senate.
In the House model, there is an average increase in partisan polarization of 0.130 for each survey year, and the model explains over eighty percent of the variance in elite partisan polarization. The story is similar for the Senate model. The per-unit increase in partisan polarization is 0.097, slightly less than that in the House model. And the model performance is equally strong: 77.9% of the variance in elite partisan polarization in the Senate is explained by partisan polarization at the mass level. What this shows is that there has been a substantial increase in partisan polarization at the mass level. Republican identifiers have become much more conservative. Democratic identifiers have become much more liberal. There has been a concomitant increase in partisan polarization at the elite level. Republican legislators have become much more conservative. Democratic legislators have become much more liberal since the 1970’s. As Figure 9.4 and the models in Table 9.5 indicate, there is a strong relationship between mass and elite partisan polarization both in direction (upward) and in magnitude. While the direct evidence of increasing intractable policy conflict and a related decline in the probability of compromise on major policy changes to the status quo presented here is anecdotal, this strong relationship between mass and elite polarization is very suggestive that the state of affairs in American politics today is closer to that depicted in Figure 9.2 rather than that shown in Figure 9.1. Republican and Democratic Presidents are more likely to play to their base constituencies, and propose more extreme policy alternatives. The legislators tasked with crafting these proposals into actual laws are less inclined to compromise, less inclined to acquiesce to moderate alternatives, and more likely to take a ‘this change or no change’ stance. The prospects of party-line policies increase the chance of a ‘no policy change’ outcome, as it exposes vulnerable party members in the next election cycle. The result: intractable conflict, party-line voting, and increasingly intense opposition from the party out of power. Whether this is the case or not, it is clear that the last thirty years plus has witnessed a remarkable and significant increase in the ideological consistency of and ideological distance between the Republican and Democratic parties in the mass public and at the elite level. Partisan Polarization on Abortion Having established partisan polarization in the ideological dimension, I turn here to the issue dimension, and specifically one of the most important issues for the culture wars: abortion. The abortion issue is the sine qua non of the culture wars. Increasing partisan polarization on the abortion issue would be further evidence that the parties are increasingly motivated on the social issue dimension and that public debate on social issues and the public policy process is increasingly structured by partisan competition. I report the group polarization measure for partisanship on the abortion issue along with the decomposition of the measure into party ID categories in Table 9.6. Table 9.6 shows significant partisan polarization on the abortion issue, mostly due to the polarization of Republicans and Democrats on the issue. In 1972, there was little to no abortion polarization (0.015). By 1990, each of the individual category contributions to polarization on abortion by the Republicans and Democrats would exceed the total polarization on abortion in 1972. Furthermore, the maximum polarization contribution for Republicans (0.104) and Democrats (0.124) is Table 9.6: Party ID – Decomposition of Partisan Group Polarization on abortion, 1972-2004
the last year in the time series, 2004. This suggests that not only has polarization on abortion significantly increased over the course of the time series, but that partisan polarization on abortion is an ongoing contemporary phenomenon. Figure 9.5 plots the total partisan polarization from 1972 to 2004. Partisan polarization throughout the 1970’s and well in to the 1980’s was relatively flat. From 1988 to 1990, however, there is a substantial spike in partisan polarization on abortion. Abortion polarization goes from 0.012 to 0.086—over a 700% increase in partisan polarization—in these two years. Note that this time period synchs up well with the culture wars thesis. Except for 1994, which witnessed a slight dip in partisan polarization on abortion, from 1988 forward partisan polarization has monotonically increased biannually until the end of the time series in 2004. Hence, the global maximum for partisan polarization on abortion is in 2004 (0.238). Interestingly, while we might have expected partisan polarization to stall-out given the return of foreign policy to prominence as one of the major political fault line in American politics, the increase in partisan polarization during the first decade of the 21st century is comparable with the partisan polarization from the 1990’s. Since the 1970’s there has been a dramatic increase in partisan polarization on abortion. The decomposition of partisan polarization by party identification category is reported in Figure 9.6. The dramatic increase in partisan polarization on abortion is apparent, with the two major parties as the biggest contributors to partisan polarization. A particularly interesting finding in the decomposition figure is the fact that it is not the case that Republicans and Democrats were the two biggest contributors to partisan polarization since the 1970’s. Up until 1988, during the low-polarization period, it was sometimes the case that the Independents outstripped either the Republicans or Democrats in contribution to partisan polarization. Indeed, in 1980, 1984, 1986, and 1988, the Independents were larger contributors to partisan polarization than the Republicans. In 1988, though it is a low polarization survey year, the Independents were the largest contributors to partisan polarization
on abortion. This leads us to 1990, the most unusual data point in the entire series, perhaps in all of the polarization analyses. 1990 not only witnessed a dramatic increase in partisan polarization across the board, but it was the Independents that were responsible for the largest contribution to this polarization. While interesting, there is no obvious reason this would be the case. Though, as noted in Chapter 7, this was in the wake of the 1989 Webster decision. After 1990, partisan polarization exhibits trends in the expected direction. By 1992, the ‘normal’ order, with Republicans and Democrats the two primary contributors to partisan polarization, had reasserted itself. Table 9.7 reports the decomposition and total partisan polarization trend models for abortion from 1972 to 2004. There is strong evidence of partisan polarization overall (R2 = .807) with an average increase of 0.007 in polarization on abortion for every survey year. The Republican and Democrat identifiers have identical parameter estimates and standard errors, though the Democratic model explains slightly more of the variation in partisan polarization (R2 = .784) over the Republican model (R2 = .745). Independents contribute to the trend in increasing partisan polarization, but not as much as the
two major parties (0.001). The overall fit is not nearly as good, explaining less than 50% of the variance in the dependent variable (R2 = .455). As I noted earlier, the trend in partisan polarization for the individual partisan categories reflects a separation of the two major political parties that began in 1990 and has steadily advanced through to 2004. The percent contributions of each partisan category for the weighted and unweighted group polarization measure can be found in Tables 9.8 and 9.9. The maximum contribution for Republicans was 52% in 1988 for the weighted polarization measure. The unweighted maximum for Republicans was in 1992, where Republicans accounted for just below 50% of the percent contribution to partisan polarization on abortion. The low point came on the weighted contribution was, unsurprisingly, in 1976. Not only was this prior to the Christian Right movement to influence the Republican Party, but it was also the first post-Watergate presidential election year (Carter), depressing the number of Republican identifiers in the sample. That said, the weighted measure downgrades the Independent category contribution much more significantly in 1976 than for the Republicans, cutting their contribution by two thirds relative to the unweighted Independent contribution. What little partisan polarization there was in the 1970’s is mostly attributable to the Democrats, again, given their size within the population. The maximum and near maximum contributions for the Democratic identifiers come in 1976, 1980, and 1982 respectively. The Democrats averaged about 68% of the contribution to partisan polarization for those years. As was the case with partisan polarization on the ideological dimension, weighting the partisan polarization measure on abortion substantially affects the contribution of Independents to polarization. In Table 9.8, the contribution of Independents to polarization, while sometimes eclipsing one of the major parties, is never the top categorical contributor to partisan polarization. This is not true of the unweighted contribution of Independents. From 1986 to 1990, the Independent category accounted for nearly half of the partisan polarization on abortion. Table 9.8: Percent Contribution & Mean Deviation Weighted Partisan G.P. on Ideology, 1974-2004
Table 9.9: Percent Contribution & Mean Deviation Unweighted Partisan G.P. on Ideology, 1974-2004
I estimate linear trend models for the mean deviation in the percent contribution of each partisan category to the polarization on abortion (Table 9.10). In assessing the linear polarization trends, I omit the 1972 data point for both empirical and theoretical reasons. Empirically, the 1972 polarization data point was an unusual outlier. While one should be cautious in eliminating data points, there is good reason to exclude the 1972 data on abortion. Slope estimates of small-N regressions are particularly susceptible to the influence of outliers. The ANES used a split-sample in 1972, and as such the abortion question was asked of only half of the 1972 sample. Further sub-setting the data by partisanship may have resulted in unrepresentative estimates of the partisan positions on abortion. Theoretically, 1972 preceded the 1973 Roe v. Wade USSC decision that sparked the cultural battle over abortion. By excluding 1972, the trend depicts the percent contribution for the categories (mean deviation) to abortion polarization since that landmark decision. For the weighted polarization measures, the linear models reveal interesting trends in the contribution of Republican identifiers to the abortion issue. There is a statistically significant positive coefficient for the Republican model, indicating that Republican contribution to abortion polarization has increased since the 1970’s. For every survey year, there is a 0.596 increase in the percent contribution for Republicans to abortion polarization. This model accounts for about 25% of the variation in percent contribution for the time series. While Republican identification has increased since the 1970’s, this trend is not exclusively a consequence of that fact. Republican contribution to the abortion polarization measure increased in the unweighted model as well (0.504). Interestingly Democrats, when you account for their declining adherents since the 1970’s up until 2004, evidence no significant trend in their contribution to abortion polarization. This is also the case when it comes to the unweighted model (R2 = .003). Substantively speaking and ignoring the lack of statistical significance for the Democratic coefficient, there is clearly a much smaller increase in magnitude when compared to that of Republicans. It seems clear that Republicans have contributed more to abortion polarization than Democrats have since the 1970’s. Irrespective of whether we weight the groups by size, Republicans evidence a significant increasing contribution to abortion polarization while the Democrats do not. Those who suggest that Republicans were ‘captured’ by the Christian Right may have exaggerated their case, but clearly the influence of conservative religious citizens increasingly identifying with and influencing the Republican Party has had an effect on their collective position on abortion. For Independents, they have contributed less to abortion polarization since the 1970’s in both the weighted (-0.314) and unweighted regression (-0.564) models. Recall that Independents were significant contributors to abortion polarization in the late 1980’s and early 1990’s but after 1992 there is a sharp decline in the contribution of Independents, and this trend remains constant through 2004. The Table 9.10: Weighted & Unweighted Mean Deviation Trend Regressions of PID GP on Abortion, 1974-2004
regression coefficients for Independents in Table 9.10 capture this change. The weighed model explains just over 25% of the variation in percent contribution, while the unweighted model accounts for 22% of the change in the Independent contribution to abortion polarization. An Example of Partisan Polarization on Abortion and Public Policy: the PBA Ban Unlike with ideology, there are no relatively objective measures of the policy outputs for Congress on abortion or the abortion policy positions for congresspersons for the full time series. While interest group ratings could serve as a proxy for abortion positions for congress, I do not include an analysis with that data here (though it is an avenue of future research). What can be done here is to look at a suggestive example of the partisan polarization on abortion and the effect this has had on abortion public policy. Figure 9.7: 2003 ABC News Opinion Poll - Public Opinion on Abortion by Situation67 
In 2003, an ABC News opinion poll attempted to assess the nature of abortion opinion across a range of possible situations where a woman might seek an abortion. It included the normal abortion ‘exceptions’ of life, health, rape, and incest, but it also included newer abortion issues such as abortions of physically impaired babies and partial birth abortion. Partial birth abortion is an abortion procedure involving a partial delivery of the fetus in order to complete the abortion, hence the term. It has been highly controversial since it became a target of the abortion debate, with the Partial Birth Abortion Ban Act of 2005 having termed the procedure “gruesome and inhumane.” Whatever the merits of the argument, as the 2003 ABC News poll illustrates, the American public is aligned in near-consensus against it. Nearly 70% of respondents to the poll (69%) said they believed the procedure should be illegal. Only 21% reported that it should be legal. While the public has long been opposed to abortions in “unwanted child” situations, that position is relatively conflictual, with over 55% of the public aligned in opposition but 41% of the public in favor of permitting abortions in that situation. That is an over 10% difference on the ‘opposed’ side and an over 20% difference between respondents who believe partial birth abortion should be legal and those who support the legality of ending unwanted pregnancies. In a depolarized, non-culture war political environment, we would not anticipate a great deal of controversy associated with the PBA ban. With the public solidly behind the ban, a major controversy over its enactment should be unlikely in a non-partisan, depolarized abortion issue dimension. As Table 9.11 illustrates, however, this is not the case. The table gives the partisan breakdown and caucus percentage for the 2005 vote on the Partial Birth Abortion Ban Act. It was voted on in both the House and the Senate, with the Harkin Amendment the only substantive difference between the two bills. The Harkin Amendment expressed support for the Roe v. Wade decision (which also enjoys majority support in polling). That amendment was eliminated in conference. That in itself is suggestive of the Figure 9.2 scenario, as Republicans rejected a Democratic amendment which might have induced more bipartisan support for the bill.
The vote illustrates clearly that even on a bill which reflects a strong, consensus position for the American people, the behavior of Republicans and Democrats in Congress reflects the evidence of strong polarization on the issue between their constituencies that this chapter has demonstrated. In the House 98% of Republicans supported the ban and 94% in the Senate did likewise. While this outstrips the level of public opposition, at least as was evident in the ABC poll from two years prior, it is fair to say that 100% of rational congresspersons should follow where 70% of the American public leads. However, that is not the case with the Democrats. Strong majorities of the Democratic caucuses in both the House and the Senate opposed the ban. While the Democratic caucus was certainly more split than the Republicans (with over 30% defections in both houses), the fact over 65% of all the Democrats in Congress voted against the ban is a testament to extent of partisan polarization on Abortion. While some might attribute this to the influence of pro-choice interest groups such as NARAL, the evidence in this chapter presents a more electorally-plausible, constituent-centered explanation. Democrats opposed the ban because their constituents tilt strongly towards the Pro-Choice position. Republicans likewise on pro-life issues, though in this case they also had a majority of the public to rely upon in justifying their votes. While one shouldn’t go too far in over-interpreting a single vote from Congress or a single poll of the public on partial birth abortion, the vote and the poll are consistent with partisan polarization on abortion and consistent with representatives to Congress responding to that polarization in the policy process. One way to think about it is to do so in terms of a counterfactual: how does this abortion vote comport with what we could reasonably expect if the Republicans and Democrats in Congress were responding to the ‘average’ voter and the strong consensus in the public against partial birth abortion or is it inconsistent with it? I would argue we would expect a near consensus on a partial birth abortion ban in the absence of partisan polarization on abortion. Since the parties in the mass public have diverged on abortion, we see this reflected in the partisan polarization on the PBA vote. Conclusion In this chapter I present compelling evidence of significant partisan polarization in public attitudes on the ideological dimension. I tie that ideological polarization at the mass electorate level directly to the voting behavior of members of Congress, a part of the American political elite. The relationship between partisan polarization on ideology at the mass level and partisan polarization level is strong, and the correspondence of polarization on ideology between the mass and elite levels is quite striking. I explore the nature of this relationship in terms of which way the causal arrow points in Chapter 12. In terms of the contributions of identifiers of either party on ideological polarization, I show that the Democrats have had an increasing role in the polarization of ideology, irrespective of the number of Democratic identifiers in the electorate. At the same time, Republicans had a corresponding decline in their contribution to ideological polarization. However, both of these trends are mediated by group size, as weighting the measures by group size revealed no significant trend for either Republicans or Democrats. The significant trend in the weighted group polarization measure of party identifiers on ideology was for the Independents, evidencing a declining contribution of polarization. As noted earlier, Independents have declined as a portion of the population since the 1970’s. Interestingly, Republican and Democratic contribution to partisan polarization on ideology is directly related to the party of the presidential administration. When a Democrat is president, Republicans contribute more to partisan polarization on the ideological dimension. When a Republican is president, it is the Democrats in the mass public who shift to the extremes on ideology. This reaction to the presidential administration suggests that, for those citizens who identify with one of the two major parties, dissatisfaction with their party being out of power tends to induce an aggregate move towards the ideological extremes. On abortion, as with ideology, I find strong evidence of partisan polarization in the American public. While the 1970’s and most of the 1980’s exhibited little partisan polarization on abortion, with the contribution of both parties to polarization sometimes eclipsed by that of Independents, this all changes in the latter half of the 1980’s. From 1988 forward, I find a substantial increase in partisan polarization on abortion that increases nearly monotonically until the end of the time series in 2004. Both Republican and Democratic identifiers were strong contributors to this polarization trend, though Republicans outstripped Democrats in percent contribution to abortion polarization in both the weighted and unweighted group polarization trend models. The case of the Partial Birth Abortion Ban Act of 2005 suggests this polarization at the level of the mass electorate has translated into polarized partisan behavior in Congress. Despite a consensus in the American public against partial birth abortion, over 65% of Democrats in both houses voted against the ban. Download 4.38 Mb. Share with your friends:
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
