Arkansas Tech University The Culture Wars & Political Polarization in Perspective
Download 4.38 Mb.
|
Status Quo on Defense Spending Status Quo on Defense Spending - 3 - 2 - 1 0 1 2 3 1 9 8 0 1 9 8 1 1 9 8 2 1 9 8 3 1 9 8 4 1 9 8 5 1 9 8 6 1 9 8 7 1 9 8 8 1 9 8 9 1 9 9 0 1 9 9 1 1 9 9 2 1 9 9 3 1 9 9 4 1 9 9 5 1 9 9 6 1 9 9 7 1 9 9 8 1 9 9 9 2 0 0 0 2 0 0 1 2 0 0 2 2 0 0 3 2 0 0 4 2 0 0 5 2 0 0 6 2 0 0 7 2 0 0 8 DEFENSE SPENDING - %GDP (Z) DEFENSE MEAN (Z) DEFENSE BIMODALITY KURT ( - Z)
For many liberal activists, opposing the war was really about opposing George W. Bush. When Bush disappeared, so did their anti-war passion. - Byron York, 2009
The third possible explanation for the variance in polarization on defense spending and foreign policy are the exogenous foreign policy-related events that are, in many respects, independent of the policy-making process, yet at the same time spur (even dictate) foreign policy decisions by government officials and defense policy actors. The Cold War dates back to the 1950’s and thus any president taking office had to grapple with the threat of the Soviet Union, the bipolar world of the Superpowers, and policies such as détente, which enjoyed a status quo position in the 1980’s. The end of the Soviet Union is a good example. It is an exogenous event (certainly from the perspective of the Bush administration, whether you credit Reagan Era foreign-policy with the collapse or not), and thus any president in office when it happens would have to respond to it. However, that response is not preordained: a range of policy choices is available to the administration. To put it another way, defense spending was bound to decrease following the end of the Cold War no matter who was the president, but by how much, and where American intervention in the international system would happen next, is very much within the purview of presidential choice. Likewise, defense spending was bound to increase in the wake of September 11th. And a war in Afghanistan against the Taliban probably would have happened had there been a President Gore in the White House. But the war in Iraq, and the foreign policy philosophy which demanded it, may not have been. This suggests an interactive effect between the party of the president (a proxy for foreign policy philosophy given the differing positions of the parties on military intervention and how the parties are perceived on defense by the mass public) and foreign policy events.
* significant at .10 level ** significant at .05 level ***significant at .01 level Table 7.2 reports a simple regression with the z-scores for actual defense spending (% GDP) as the predictor of the z-score for the mean position on defense spending in the general public as well as the z-score for bimodality in defense spending opinion levels. Equation 7.3: Simple Regression of Actual Defense Spending on D.S. Public Opinion
For the full time series, the public is almost perfectly middle-of-the-road. The mean public opinion on defense spending levels for the full time series (mean of the mean) is 4.074, reflecting a slight preference for more defense spending over the time span. But as can be seen in Figure 7.11, this average masks substantively important variance in the average public position on defense spending. During the 1980’s during the Reagan administration and at the height of the Cold War, the preference on defense spending is below the mean, indicating a general preference for less spending (relative to the defense spending that was going on). During the 1990’s, the public preferred more defense spending, on average, than was the case with the Clinton administration’s deep defense spending cuts from 1994 through the end of his presidency. And while the preference for more defense spending spiked after September 11th, as the war in Iraq drug on that average preference for defense spending regressed towards the mean defense spending preference level (Figure 7.11). The coefficient for actual defense spending in Table 7.2 suggests this inverse relationship is significant and substantively important. For a single unit increase in actual defense spending, preference for defense spending in the public declines -0.478 Z-score units. However, this model has a relatively weak fit to the data, with only about 30% of the variance in average defense spending public opinion explained by actual defense spending levels (% GDP). There is a strong central tendency in American public opinion on defense spending. Republican administrations appear to have endorsed defense spending levels in excess of that preferred by the public, while Democratic administrations have implemented deeper defense spending cuts above and beyond what is preferred by the public. The relationship between actual defense spending and bimodality in public opinion on defense spending is positive and much stronger than that of defense spending and public opinion on defense spending. For every single unit increase in actual defense spending, there is a 0.859 unit increase in defense spending opinion polarization (kurtosis), and this model explains over seventy percent of the variance in defense spending public opinion bimodality. To test all three of the theories on defense spending public opinion, I create dummy variables for the partisanship of the presidential administration and the foreign policy period. Table 7.3 reports the structure of the dummy variables. The presidential administration partisanship dummy is a simple binary categorical variable with the value of one for a Republican president and zero for a Democratic president. The three level variable for foreign policy period gives a -1 for the New World Order period of foreign policy extending from the fall of the Berlin Wall in 1989 to the September 11th attacks in 2001. The Cold War period (value = 0) extends for all years prior to the fall of the Berlin Wall and the War on Terror period (value = 1) extends from the September 11th attacks through to the present. TABLE 7.3: Coding Scheme for Foreign Policy Period & Presidential Party Variable
Table 7.4 adds the measure for presidential administration partisanship to the actual defense spending model, examining the variance in defense spending public opinion in a multivariate analysis. Equation 7.4: Multiple Regression - Defense Spending Public Opinion Model 
The model on average public opinion on defense spending levels is greatly improved by the inclusion of presidential administration party in the regression. There is a fifty percent improvement in the overall fit of the model by including party as a predictor variable, as the multivariate model explains close to half of the variance in average defense spending opinion (R2 = .467). While actual defense spending has an independent, significant affect on average defense spending public opinion, the multivariate analysis makes it clear that public preference on defense spending levels is highly dependent upon what party controls the White House. This could be a function of opposition party portrayal of the president as more extreme on defense spending than he actually is (I address partisan perceptions of the opposition in Chapter 7), or it could be due to institutional and partisan constraints on presidential policymaking. Partisans have incentives to portray an opposition president as more extreme than their actual policy positions may warrant. And there are significant institutional roadblocks in the American system of separated powers that can limit the extent to which a president can move the status quo. As reported TABLE 7.4: D. M. Regressing Defense Spending & Party of Presidential Admin on D.S. Public Opinion.
* significant at .10 level ** significant at .05 level ***significant at .01 level by Vin Weber, President Clinton had promised “dramatic” defense spending cuts as he took office in 1992 (Weber 1993). However, the deep cuts in Reagan/Bush spending didn’t materialize in his first budget. Defense Secretary Lee Aspin declared, “"What we're doing is kind of treading water." Why did Clinton adopt what was described as a ‘holding budget’ rather than institute the steep cuts he preferred: a $122 billion dollar cut over four years—twice what he had proposed on the campaign? Institutional resistance, as Colin Powell referred to the budget as “fundamentally flawed” as well as resistance from centrists within his own party, such as Senator Sam Nunn (Weber 1993). Public perception of presidential defense spending may be influenced by promises and proposals that, through the course of the policy process, become watered-down significantly in comparison to the original intentions. The bimodality model does not show as dramatic an improvement vis-à-vis the mean model with the inclusion of party as an independent variable. However, the party of the presidential administration is a significant predictor of defense spending. The sign of the coefficient is negative, indicating that there is less conflict for Republican administrations. However, this is due largely to the first Bush administration, which had low levels of conflict despite relatively high levels of defense spending. Still, it suggests that when you control for actual levels of spending the influence of partisanship is not in the expected direction. Yet, as I’ve noted, presidents influence actual defense spending levels. There may be an interactive relationship between these two variables that needs to be accounted for. Table 7.5 adds the categorical variable for foreign policy period to both models and reports coefficients, standard errors and model-fit statistics. Thus we now have a model with variables representing all three theories explaining defense spending public opinion competing in the same multivariate model: Equation 7.5: Multiple Regression - Defense Spending Public Opinion Model
* significant at .10 level ** significant at .05 level ***significant at .01 level For both the mean and kurtosis models, the foreign policy period is a significant predictor of defense spending opinion. In the mean model, the positive coefficient indicates that the American public prefers higher levels of defense spending in times of war, such as the Cold War period and the War on Terror. Again we see a significant improvement in the predictive capacity of the multivariate model, with just over sixty-six percent of the variance in average defense spending opinion explained. In the bimodality model, inclusion of foreign policy period improves the overall fit of the model to .820, meaning that over eighty percent of the variance in defense spending bimodality is explained by actual defense spending, the party of the president, and the exogenous event-determined foreign policy period. Given that the war periods have higher values than the New World Order period, it isn’t surprising to see that, as foreign policy period increases, so does the polarization in defense spending opinion. What is interesting is that all three variables have independent significant influence on defense spending opinion, controlling for the other variables in the model. Table 7.6 reports the full models for both average defense spending opinion and the bimodality of defense spending with interactions.54 Equation 7.6: Multiple Regression - Defense Spending Public Opinion Model w/ Interactions 
In the mean model, actual defense spending level continues to have a significant, negative influence on defense spending public opinion (-1.283). Even controlling for the other theoretical explanations and interactions, the public’s opinion on what the level of defense spending should be is responsive to defense spending levels. The negative coefficient indicates an inverse relationship between actual defense spending and defense spending opinion. As defense spending increases (becoming more extreme relative to the mean) for the time period, the public wants less spending on defense. As defense spending decreases (becoming more extreme relative to the mean), the public wants more spending. Partisanship (of the presidential administration) has a substantial and significant effect on average defense spending opinion independent of the other included variables (2.197). Beyond the exogenous shocks to the system and the actual defense spending levels, the party of the president is a significant factor. Specifically, citizens prefer higher defense spending during Republican administration years relative to Democratic administration years. Confirming earlier speculation, there are significant interaction effects between defense spending levels and the foreign policy period. Naturally events like the Cuban missile crisis, the Soviet
* significant at .10 level ** significant at .05 level ***significant at .01 level invasion of Afghanistan, and the terrorist attacks on September 11th, can precipitate increases in defense spending in and of themselves. Likewise, liberalization in the USSR (Perestroika), the fall of the Berlin Wall, and the collapse of the Soviet Block result in demilitarization. The combined effects of changes in actual defense spending and foreign policy events—that serves either to make their world look safer or more dangerous—influence public opinion on defense spending levels. While the overall effect of defense spending remains negative (even with the interaction), foreign policy events tends to mitigate this inverse relationship. There is also a significant interaction between the foreign policy periods and the party of the president. The positive coefficient indicates that times of war combined with Republican administrations tend to produce higher preferred defense spending levels than otherwise. Turning to the model of bimodality, levels of defense spending as a percentage of GDP tend to produce more polarization, independent of the other variables and interactions. The strong, positive coefficient (1.218) indicates that there is more bimodality in defense spending opinion coincident with higher levels of defense spending. Defense spending levels and the party of the president has an independent effect on defense spending opinion. The positive coefficient (0.387) suggests that it matters what party the president is when defense spending increases. Republicans tend to provoke more polarization when they make changes to defense spending than Democrats. Interestingly, when accounting for the interactive effect of the foreign policy period and the party of the president, partisanship is not a significant predictor of defense spending opinion. This suggests that the public is not reacting to party alone, but rather to the party of the president in the context of exogenous events. Bush’s partisanship may have engendered more conflict in defense spending than a Gore presidency would have engendered, but September 11th, and the foreign policy decisions that resulted because of it, were necessary factors. In a world in which September 11th doesn’t happen and Bush continued the minimalist foreign intervention policy he had signaled in the campaign, it is likely that the divisions in society over defense spending wouldn’t have become as polarized as they did in response to the wars in Iraq and Afghanistan. There is significant and substantively important polarization of defense policy opinion, both in terms of the average public-preferred level of defense spending and the conflict over defense spending (kurtosis). The poorly performing linear trend model on bimodality for defense spending opinion masked an important relationship between defense spending opinion, actual levels of defense spending, partisanship, and the defense-related events of the past three decades. Clearly the partisanship of the presidential administration, the primary actor in American foreign policy, affects the preferred defense spending levels of the American public. Partisanship influences the preferred levels of defense spending, and presidents are the most significant players in determining the defense budget. But this effect is mediated by foreign policy events as well as institutional and political constraints. There is an inverse Figure 7.12: Bimodality Trends in Distributions of Mass Opinion on Social Issues, 1972-2008 Download 4.38 Mb. Share with your friends:
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
