SlideShare a Scribd company logo
1
Using Indirect Policy Feedback to
Understand Public Opinion for
Improving and Protecting the
Environment
Jeremy Craig Green
Anthony Leiserowitz
Arnab Pal
Abstract
In this study, we measured the extent to which the political party of the President of the
United States influences the electorate’s willingness to pay for improving and protecting
the environment. We used General Social Survey (GSS) data to develop discrete choice
models that dynamically gauged collective public opinion on environmental spending
from 1973-2010. This learning model of policy feedback monitors public response to
policies as those policies shift over time. Our model identified a strong, linear
relationship between consecutive years of Republican presidency and the perceived
probability of spending too little to improve and protect the environment. During
Democratic administrations, about 57% of individuals consider environmental spending
to be too little. This shift accumulates over time, and after 10 years of consecutive
Republican administrations, that number rises to 73%. Despite this increased feedback
concerning comparably low levels of environmental spending, Republican
administrations continued to decrease spending relative to Democrats, and after 10
consecutive years in office, Republicans had decreased spending levels by nearly $40
billion (adjusted for inflation, in 2010 dollars). Individual characteristics (party
2
affiliation, political ideology, income, and education categories) are also assessed using
discrete choice methods.
Introduction
Policy feedback theory holds that as policy actions in one direction or the other increase,
public opinion will shift against the given direction of policy, and politicians may then, in
turn, respond to opinion by moving their policies in the other direction. This theory
requires individuals to have knowledge of these policies and to respond to public opinion
surveys accordingly, in order to provide feedback expressing their policy preferences.
Politicians, on the other hand, have a responsibility to monitor public opinion and
respond to it, although the extent to which this happens continues to be a topic of deep
debate among political scientists. This paper provides a fresh perspective on these
streams of literature by developing a policy-learning model of public opinion, where
responses to public opinion surveys shift over time in association with consecutive years
of Republican presidency. We focus on public opinion for willingness to pay for
improving and protecting the environment, a domain that has grown increasingly partisan
over the past 1-2 decades and importantly, is a topic for which individuals may have
relatively little direct, personal experiences with the major policy issues (heat waves
might be a noticeably exception), and may thus be likely to base their attitudes on some
secondary source of information, such as the political party of the President. Attitudes
toward health and military spending by the government are also evaluated in order to
compare the extent to which environmental attitudes may be more or less partisan than
other policy domains.
3
In this essay, we develop and test a learning model of policy feedback, where the public
responds to policies as the policies shift over time. The main innovation of our approach,
in comparison to that taken in previous studies, is that we examine more dynamically
accumulating changes in public opinion that occur over the course of continuous,
consecutive years of Republican presidency. Using data from the General Social Survey
(GSS) from 1973 to 2010, discrete choice models were used to test for shifts in the
willingness to pay (WTP) distribution for federal spending on the environment and
natural resources, as well as for other policy domains. In particular, we tested for a linear
relationship between consecutive years of Republican presidency and the perceived
probability of spending too little, about right, or too much to improve and protect the
environment. To quantify the extent to which individual survey responses represent
stereotypes of Republicans as presidents with low levels of environmental spending, or
whether Republicans actually do spend less than Democrats over consecutive years of
administration, we tested for a corresponding relationship between consecutive years of
Republican presidency and actual levels of spending, compared to an empirically robust
average of spending during Democratic administrations.
Review of the Literature
Previous theoretical and empirical descriptions of the relationships between public policy
and public opinion offer a variety of explanations for changes in public attitudes over
time. Individual attitudes reported in public opinion surveys seem to best reflect changes
in public policy if individuals are paying particular attention to a given issue and have
4
meaningful beliefs concerning a given topic, but they might be less meaningful in cases
where individuals do not have specific knowledge of a topic at hand. When asked about
their preferences for government spending to improve and protect the environment,
individuals may not have meaningful beliefs about government spending insofar as they
do not know the actual levels of government spending. Although survey respondents
may not have access to this primary information about government expenditures, their
beliefs might still be meaningful if they can access some type of “secondary” information
to help them formulate accurate assessments of a concept on which they have no primary
information.
Stimson (1991) notes that long-term policy preferences across different issues are
connected and refers to these long-term changes as “moods.” Wlezien (1995) describes a
negative feedback loop between policy outputs and public inputs that occurs when the
public’s preferences for spending on policy are inversely related by a government’s
spending decisions. In a Democratic polity, information gaps may be narrowed through
the process of negative feedback (Stimson, 2004). According to this theory, if the current
policy differs from the desired policy, the public will send a message to policy makers to
modify the policy to the public’s preference (i.e., spending level). For example, if a
government expenditures increases (decreases) for environment and natural resources
spending beyond a desired level (for the public), public support for expenditure on
environment and natural resource spending will adjust by decreasing (increasing),
suggesting that the public processes information on budgetary policy for these issues
accurately.
5
In the case of government spending on environment and natural resources, whether or not
public opinion responds to government policy seems to depend on how government
policy is conceptualized theoretically and empirically. Some research has suggested that
public opinion does not respond to individually delineated social expenditures (such as
health, education, environment, cities) by the government (i.e., the information gap
persists), which might suggest that the public cannot discriminate specific social policy
change with regards to changes in appropriations (Wlezien, 1995). By not responding to
government appropriation levels for specific social issues, we could potentially fail to
participate in Wlezien’s negative feedback loop with regards to specific social policy
related to environmental protection. Importantly, however, even if we are unable to
provide an accurate response for how much we are willing to pay for environmental
protection through signals created by environmental conditions and/or policy
appropriations, we may instead be able to rely on secondary information.
Secondary Information
According to Page and Shapiro (1992), individuals may adjust their policy preferences
based on new policy pertinent facts (i.e., changes in federal appropriations). However,
individuals are more likely to respond to this new information “using cognitive shortcuts
or rules of thumb, such as reliance upon trusted delegates or reference figures (friends,
interest groups, experts, and political leaders) to do the political reasoning for them and to
provide guidance.” (Page and Shapiro, 1992) This idea is supported by Gomez and
Wilson (2001) in their research regarding the economic judgments of less sophisticated
6
(or knowledgeable) voters, who attribute “responsibilities for outcomes to the most
obvious actor in the relevant sphere.” In regard to the national economy, this secondary
source would be the President. These less sophisticated voters attribute the national
economic condition more so to the President than to other economic forces and policies
that cannot be controlled by the President. Gomez and J Wilson (2001) also note that
more sophisticated voters are less likely to make this direct attribution and more likely to
evaluate the national economic condition based on multiple variables and sources of
information.
Although secondary information may be more cognitively accessible to the average
individual than primary information, it may also be less accurate. Stimson (2004)
proposes that the public does receive feedback in terms of how the environment is doing
overall (i.e., Is the environment healthy or unhealthy?). Unfortunately, the public’s
knowledge of environmental issues has historically been unsophisticated and limited
(Arcury, 1987). This gap in the public’s knowledge has persisted with newer
environmental concerns, such as climate change (Leiserowitz, 2010), which is an issue
that is difficult for the public to understand and relate to in their daily lives (Lorenzoni,
2006).
The political party of elected officials might be a particularly salient source of secondary
information that individuals might rely on when responding to questions about federal
spending on natural resources and the environment. Partisan political leaders have taken
positions that coincide with individual voting partisans, as Republicans and conservatives
7
are less likely to support further government efforts to protect the environment (Konisky,
D. M., Milyo, J., & Richardson, L., 2008). The history of partisanship in federal
environmental spending and legislation dates back to the origins of environmental issues.
Environmental legislation has been partisan, as noted by congressional Republicans in the
1970’s, who were significantly less likely than congressional Democrats to vote for the
environmental legislation passed by the then-Republican-led Nixon and Ford
administrations (Dunlap & Allen, 1976). Within both Congress and the White House,
partisan differences in support for environmental protection have only strengthened over
time, as Republican congressional leaders and the administrations of Presidents Reagan
and George W. Bush have championed anti-environmental protection policy positions,
while Democratic administrations and congressional leaders have more homogenously
moved towards more pro-environmental protection policy stances (Dunlap, 2001).
The political party of the President may serve as the most effective secondary
information source for spending on environmental protection. A 2012 PEW Research
Center survey noted that more Americans could identify the political party of former
presidents Ronald Regan, Bill Clinton, and John F. Kennedy (85%, 84%, 78%
respectively) than the political parties of current congressional leaders Nancy Pelosi and
John Boehner (61% and 55% respectively) (Kohut et. al). Carpini and Keeter (1993)
conducted an analysis of the 1989-1991 National Election Surveys to determine the
factual political knowledge of the public by examining various survey questions related
to politics and government policy. The surveys asked participants to identify political
leaders and their political parties, recognize trends in government appropriations for
8
defense and social policy, and answer various civics questions. An analysis of the surveys
showed that the party identification of a former President (Nixon, p=.78) was a more
significant predictor of overall factual political knowledge than the current party
identification of the congressional (House p=.68 & Senate p=.55) majorities. Both forms
of party identification are more significant in predicting political factual knowledge than
recognizing trends in government spending for both defense (p=.28) and education
spending (p=.24).
Public Engagement
Other theoretical and empirical literature provides alternative explanations for changes in
public opinion over time. It is very difficult for the public to answer a question on how
much our government should spend to improve and protect a set of issues, as a proper
answer requires knowledge that can only come from full-time attention to a specific issue
(Stimson, 2004). The notion that respondents do not have meaningful attitudes or beliefs
has been widely contested by academics, e.g., Feldman (1989) & Page and Shapiro
(1992). For example, Converse’s (1964) study suggests that large portions of the public
may not have meaningful knowledge about survey questions, so they instead offer
whatever response they think may please the interviewers. Whether or not individuals
are able to provide meaningful responses to a given survey question depends on the
question, as well as the policy domain of import.
Historically, public knowledge of environmental issues has been limited. Arcury and
Johnson (1987) brought the issue of public environmental knowledge to the forefront by
9
using a statewide survey to determine that public knowledge is low and also directly
correlates with education, income, and sex. The level of public environmental knowledge
has not changed even as environmental issues have. Yale University’s Americans’
Knowledge of Climate Change (2010) survey report noted that over half of all
respondents failed (answered less than 60% of 81 graded questions correctly) the public
knowledge of climate change survey, thus confirming that a majority of Americans do
not have a fully developed understanding of climate change. Hunter and Rinner (2004)
explore this public knowledge issue one step further by analyzing the relationship
between public concern for and knowledge of environmental issues. The study uses a
Colorado state survey that measures public concern for species diversity using the New
Ecological Paradigm (NEP) scale, and concludes that those with greater knowledge of the
species are not significantly likely to be more concerned with the species itself.
Policy Feedback
Scholars have long examined associations between public opinion and public policy
action from governments (Eichenberg, R. C. and R. Stoll 2003; Erikson, MacKuen, and
Stimson 2002). Page and Shapiro (1983) examine public opinion and policy data from
1935 to 1979 to conclude that opinion usually moves before policy and is, thus, an
important factor for policy change. When examining public opinion’s impacts on specific
public policy (defense spending), Hartley, T. and B. Russett (1992) find evidence that
public opinion influences military spending. If public sentiment for more military
spending rises, actual government spending for military spending will increase (and vice
versa). Stimson, J. A., M. B. MacKuen, et al. (1994) used empirical evidence to support
10
an alternative conclusion, where in the long run, shifts in public opinion are tied to shifts
in public policy. Within this relationship where policy is the leading indicator, Wlezien
(1995) noted a negative feedback loop in regard to issues of defense, welfare, and general
social spending (i.e., appropriations for environment, health, welfare, and big cities).
When examining public spending preferences for defense and general social spending
between 1977 and 1991, Wlezien (1995) concludes that the public becomes more (less)
supportive of government spending as government appropriations decrease (increase).
However, the public’s response to military spending is much stronger than its response to
social spending. A 1% increase in appropriations for social programs led to a .18%
decrease in public support, while an equal increase in defense appropriations led to a
decrease in public support (2.7%).
An examination of opinion policy dynamics in Canada by Soroka, and Wlezien
(2004) note that the public in Canada responded to changes in public spending across
various domains, though the response is less pronounced than in the US. Soroka, S. N.
and C. Wlezien (2005) conduct a similar analysis of opinion representation and policy
feedback in the UK and note through empirical evidence that the public adjusts its
preferences for various spending domains in response to spending itself, thereby
providing additional evidence for the negative feedback phenomenon present in both the
U.S. and Canada. The study notes that the negative feedback effect in the U.K. is more
pronounced than the effect in the United States. Wlezien (2004) empirically shows that
the United States government adjusts to this negative feedback loop by changing defense,
welfare, and general social policy spending to levels that are desired by the public. This
policy representation is less pronounced in the U.K. and Canada, where policy makers are
11
less representative of their constituencies under their respective parliamentary systems
than they would be under the presidential system present in the United States (Soroka, S.
N. and C. Wlezien 2004; Soroka, S. N. and C. Wlezien 2005).
Hypothesis
Based on Stimson’s theory of negative feedback and our understanding of secondary
source information, we expect a positive linear relationship between consecutive years of
Republican presidency and the perceived probability of spending too little to improve and
protect the environment, even when adjusting for a non-partisan.
Null Hypothesis 1: When analyzing the period from 1973-2010 in the GSS survey, the
perceived probability of spending too little on the environment in the US electorate,
adjusting for a non-partisan, will not significantly and linearly increase with consecutive
years of a Republican presidency.
Methods
Relationships between consecutive years of Republican presidents, willingness to pay for
improving and protecting the environment, and individual characteristics (party
affiliation, political ideology, and income and education categories) were assessed using
discrete choice methods. To test for policy learning—the accumulation of policy
feedback over time—we looked for a linear relationship between consecutive years of
Republican presidency and willingness to pay for improving and protecting the
12
environment, conditional on a Republican president in office. Choices were modeled
using an underlying regression function:
(1)
The expected WTP for an individual surveyed in year is expressed in equation (1) as
a probability function (F) of the number of continuous, consecutive years of Republican
presidency, a binary indicator for Republican presidency, and interactions between each
of these explanatory variables and individual characteristics: 10 quantiles of family
income in constant dollars as approximated by the GSS, 5 education categories, 7
categories of political party affiliations, and 7 categories of political ideology.
Substantive implications of the underlying model were assessed by examining the slope
of the linear relationship between WTP and consecutive years of Republican presidency
with the relative change in WTP from 1 year after a Republican takes office to 10
consecutive years of Republican presidency (9 years after baseline). The slope of the
linear relationship may be formally defined as an average marginal effect (AME), or the
partial derivatives of the WTP function for each individual , separately, and averaged
over the estimation sample, . The relative change from a baseline WTP was estimated
from the percent change from a baseline probability one year after a Republican president
takes office to 10 consecutive years of Republican presidency (nine years after baseline),
conditional on a Republican president in office. Each of these implications of the model
was formally tested as follows, respectively:
13
(2)
The regression specified in equation (1) was estimated using an ordered probit model
(McKelvey and Zavoina 1975)-ERROR; quantities of interest specified in equations (2)
and (3) were estimated using the observed-value approach (Hanmer and Ozan Kalkan
2012)-ERROR. Conditional marginal effects were estimated to assess the implicit and
explicit interactions in the model (Brambor 2005)-ERROR. Standard errors of interest
were clustered on survey year (Froot 1989)-ERROR, and estimated using the delta
method (Herron 1999)-ERROR.
Data and Study Sample
Data on willingness to pay to improve and protect the environment were drawn from the
General Social Survey (GSS) (McKelvey and Zavoina 1975; Smith et al. 2011). Most
importantly, this survey included a question about individual attitudes toward levels of
government spending on the environment from 1973 to 2010 for American adults ages 18
and above:
We are faced with many problems in this country, none of which can be solved
easily or inexpensively. I'm going to name some of these problems, and for each
one I'd like you to name some of these problems, and for each one I'd like you to
tell me whether you think we're spending too much money on it, too little money,
or about the right amount. Are we spending too much, too little, or about the right
amount on improving and protecting the environment? (Smith et al. 2011)
14
From 1973 to 2010, 61% of GSS respondents reported that they believed that our society
was spending too little on the environment, 30% thought that spending was about right,
and 9% said that we are spending too much. Data on the precise timing of presidencies
were pulled from Congressional Quarterly (Hanmer and Ozan Kalkan 2012; Stanley and
Niemi 2009). When presidents were both in and out of office during a single survey year,
we marked the president who was in office for most of the year. Individual characteristics
entered into the model at the highest level of detail available in the GSS and interacted
with the key explanatory variables: 7 categories of political ideology and of party
affiliation, 5 education categories, and 10 quantiles of family income in constant dollars
as approximated by the GSS from the survey categories. Data on government spending
for environment and natural resources, health, and arms and military were pulled from
the White House Office of Management and Budget (Brambor 2005; US White House
2011). Individuals who did not respond to the question (less than 1%) and those who
said that they just “don’t know” (less than 5%) were excluded from the sample, because
these two categories do not fit into the ordered levels of WTP responses.
Figures for Results
15
0.2.4.6.8
Figure 1: Perceived Probability of Spending to Improve and Protect the Environment, 1973-2010
Pr(Spending Too Little)
Pr(Spending About Right)
Pr(Spending Too Much)
1970 1980 1990 2000 2010
Survey Year
N = 28776
16
0.2.4.6.8
Figure 2: Perceived Probability of Spending to Improve and Protect the Environment, Self-Identified
Republicans (1973-2010)
Pr(Spending Too Little)
Pr(Spending About Right)
Pr(Spending Too Much)
1970 1980 1990 2000 2010
Survey Year
N = 9894
17
0.2.4.6.8
Figure 3: Perceived Probability of Spending to Improve and Protect the Environment, Self-Identified
Conservatives (1974-2010)
Pr(Spending Too Little)
Pr(Spending About Right)
Pr(Spending Too Much)
1970 1980 1990 2000 2010
Survey Year
N = 8256
18
-50-40-30-20-100
.6.65.7.75.8
Pr(SpendingTooLittle)
Figure 4: Demand for Federal Environment and Natural Resource Spending, by Consecutive
Years of Republican Presidency (compared to an average of the Carter, Clinton, and Obama
administrations)
0 2 4 6 8 10
Consecutive Years of Republican Presidency
N = 27462
Note. Perceived spending is adjusted for a nonpartisan, linear time trend; actual spending is adjusted for inflation.
19
-300-200-1000
.6.65.7.75.8
Pr(SpendingTooLittle)
Figure 5: Demand for Federal Health Spending, by Consecutive Years of Republican Presidency
(compared to an average of the Carter, Clinton, and Obama administrations)
0 2 4 6 8 10
Consecutive Years of Republican Presidency
N = 27462
Note. Perceived spending is adjusted for a nonpartisan, linear time trend; actual spending is adjusted for inflation.
20
02004006008001000
.3.35.4.45.5
Pr(SpendingTooMuch)
Figure 6: Demand for Federal Defense Spending, by Consecutive Years of Republican Presidency
(compared to an average of the Carter, Clinton, and Obama administrations)
0 2 4 6 8 10
Consecutive Years of Republican Presidency
N = 27462
Note. Perceived spending is adjusted for a nonpartisan, linear time trend; actual spending is adjusted for inflation.
Figure 7: Public Opinion and Consecutive Years of Republican Presidency, by Political Ideology (1974-2010)
21
Pr(SpendingTooLittle)
.5.6.7.8
Self-Identified Political Ideology
Liberal
Moderate
Conservative
1 10
Consecutive Years of Republican Presidency
N = 27462
Note. Adjusted for education, income, party affiliation and a nonpartisan, linear time trend.
Figure 8: Public Opinion and Consecutive Years of Republican Presidency, by Political Ideology (1974-2010)
22
Pr(SpendingTooLittle)
.65.7.75.8.85.9
Self-Identified Political Ideology
Extreme Liberal
Liberal
Slight Liberal
1 10
Consecutive Years of Republican Presidency
N = 27462
Note. Adjusted for education, income, party affiliation and a nonpartisan, linear time trend.
Figure 9: Public Opinion and Consecutive Years of Republican Presidency, by Political Ideology (1974-2010)
23
Pr(SpendingTooLittle)
.5.55.6.65.7.75
Self-Identified Political Ideology
Slight Conservative
Conservative
Extreme Conservative
1 10
Consecutive Years of Republican Presidency
N = 27462
Note. Adjusted for education, income, party affiliation and a nonpartisan, linear time trend.
Figure 10: Public Opinion and Consecutive Years of Republican Presidency, by Party Affiliation (1974-2010)
24
Pr(SpendingTooLittle)
.5.55.6.65.7.75
Self-Identified Party Affiliation
Democrat
Independent
Republican
1 10
Consecutive Years of Republican Presidency
N = 27462
Note. Adjusted for education, income, political ideology and a nonpartisan, linear time trend.
Figure 11 : Public Opinion and Consecutive Years of Republican Presidency, by Party Affiliation (1974-2010)
25
Pr(SpendingTooLittle)
.4.5.6.7.8
Self-Identified Party Affiliation
Strong Democrat
Democrat
Independent Near Democrat
Independent
Independent Near Republican
Republican
Strong Republican
1 10
Consecutive Years of Republican Presidency
N = 27462
Note. Adjusted for education, income, political ideology and a nonpartisan, linear time trend.
26
Table 1: Associations Between Republican Presidency and Demand for Federal Environment
Spending, 1973-2010
1973-2010 Sample 1974-2010 Sample
Adjustments: Unadjusted Time Trend Unadjusted Time Trend Covariates
Spending Too Little
Consecutive Years Republican 0.0160úúú 0.0132úúú 0.0189úúú 0.0162úúú 0.0155úúú
(0.00363) (0.00380) (0.00318) (0.00318) (0.00297)
% change, 1 to 10 years 23.05úúú
18.92úúú
27.93úúú
23.62úúú
22.25úúú
(5.659) (5.740) (5.081) (4.895) (4.630)
Spending About Right
Consecutive Years Republican -0.00929úúú -0.00769úúú -0.0110úúú -0.00938úúú -0.00905úúú
(0.00204) (0.00217) (0.00179) (0.00184) (0.00162)
% change, 1 to 10 years -28.19úúú -23.78úúú -32.47úúú -28.29úúú -27.35úúú
(5.579) (6.178) (4.748) (5.097) (4.546)
Spending Too Much
Consecutive Years Republican -0.00669úúú -0.00554úúú -0.00797úúú -0.00681úúú -0.00644úúú
(0.00161) (0.00164) (0.00143) (0.00137) (0.00139)
% change, 1 to 10 years -51.94úúú -45.35úúú -58.07úúú -52.28úúú -47.13úúú
(8.083) (9.640) (6.268) (7.233) (7.728)
individuals 28776 28776 27462 27462 27462
survey years 27 27 26 26 26
linear time trend X X X
covariate adjustments X
Standard errors in parentheses
ú
p < 0.05, úú
p < 0.01, úúú
p < 0.001
Note. Covariates are education, income, party affiliation, and political ideology.
27
Table 2: Associations Between Republican Presidency and Demand for Federal Environment,
Health, and Defense Spending (1974-2010)
Spending Too Little
Environment Health Defense
Consecutive Years Republican 0.0155úúú 0.00456 -0.0117úú
(0.00297) (0.00293) (0.00433)
% change, 1 to 10 years 22.25úúú 5.746 -37.18úúú
(4.630) (3.870) (11.23)
Spending About Right
Consecutive Years Republican -0.00905úúú -0.00305 -0.00181
(0.00162) (0.00195) (0.00111)
% change, 1 to 10 years -27.35úúú
-10.48 -7.837ú
(4.546) (6.323) (3.461)
Spending Too Much
Consecutive Years Republican -0.00644úúú
-0.00152 0.0135úú
(0.00139) (0.00102) (0.00514)
% change, 1 to 10 years -47.13úúú -19.06 40.40ú
(7.728) (12.36) (17.56)
individuals = 27462, survey years = 26
ú p < 0.05, úú p < 0.01, úúú p < 0.001
Note. Adjusted for education, income, party affiliation, political ideology and a nonpartisan, linear time trend.
28
Table 3: By ideology
Extreme
Liberal Liberal
Slight
Liberal Moderate
Slight
Conservative Conservative
Extreme
Conservative
Spending Too Little
Consecutive Years Republican 0.0290úúú
0.0101 0.0128úú 0.0169úúú 0.0264úúú 0.0199úúú
0.000885
(0.00576) (0.00644) (0.00393) (0.00379) (0.00511) (0.00571) (0.00779)
% change, 1 to 10 years 28.65úúú 11.47 15.91úú 24.23úúú 43.55úúú 35.29úúú 1.555
(5.113) (7.346) (4.975) (5.919) (10.32) (10.26) (13.62)
Spending About Right
Consecutive Years Republican -0.0209úúú
-0.00728 -0.00869úúú -0.0101úúú -0.0136úúú -0.00955úúú
-0.000544
(0.00375) (0.00449) (0.00248) (0.00202) (0.00193) (0.00274) (0.00374)
% change, 1 to 10 years -67.92úúú
-26.65 -28.54úúú
-29.77úúú
-39.12úúú
-27.18úúú
-2.128
(8.624) (15.30) (7.500) (5.642) (5.968) (8.168) (9.389)
Spending Too Much
Consecutive Years Republican -0.00810úúú
-0.00277 -0.00412úú
-0.00677úúú
-0.0128úúú
-0.0103úúú
-0.000341
(0.00206) (0.00196) (0.00146) (0.00181) (0.00327) (0.00302) (0.00406)
% change, 1 to 10 years -85.61úúú -41.65 -46.40úúú -51.66úúú -66.73úúú -53.27úúú -0.281
(6.582) (24.20) (11.66) (8.978) (7.769) (11.68) (27.95)
individuals = 27462, survey years = 26
ú p < 0.05, úú p < 0.01, úúú p < 0.001
Note. Adjusted for education, income, party affiliation and a nonpartisan, linear time trend.
29
Table 4: By party
Strong
Democrat Democrat
Independent
Near Democrat Independent
Independent
Near Republican Republican
Strong
Republican
Spending Too Little
Consecutive Years Republican 0.0176úúú 0.0138úúú 0.0105úú 0.0156úúú 0.0143ú 0.0201úúú 0.0172ú
(0.00344) (0.00289) (0.00324) (0.00424) (0.00605) (0.00345) (0.00696)
% change, 1 to 10 years 22.61úúú 18.89úúú 12.99úú 21.32úúú 21.73ú 31.76úúú 34.44ú
(4.826) (4.269) (4.223) (5.726) (9.539) (6.079) (15.54)
Spending About Right
Consecutive Years Republican -0.0113úúú -0.00846úúú -0.00707úúú -0.00961úúú -0.00805ú -0.0108úúú -0.00693úú
(0.00191) (0.00157) (0.00204) (0.00250) (0.00316) (0.00164) (0.00252)
% change, 1 to 10 years -35.11úúú
-25.91úúú
-23.42úúú
-29.34úúú
-23.77úú
-31.51úúú
-20.76ú
(5.377) (4.547) (6.203) (7.386) (9.171) (4.695) (8.099)
Spending Too Much
Consecutive Years Republican -0.00632úúú
-0.00530úúú
-0.00347úú
-0.00602úúú
-0.00627ú
-0.00935úúú
-0.0102ú
(0.00156) (0.00134) (0.00122) (0.00176) (0.00293) (0.00188) (0.00454)
% change, 1 to 10 years -54.72úúú -42.80úúú -36.23úúú -47.93úúú -41.85úú -54.65úúú -44.48úú
(7.888) (7.681) (10.88) (10.82) (15.62) (6.799) (14.69)
individuals = 27462, survey years = 26
ú p < 0.05, úú p < 0.01, úúú p < 0.001
Note. Adjusted for education, income, political ideology and a nonpartisan, linear time trend.
30
Results
Descriptive statistics of the estimation sample are displayed in Figure 1. This figure
shows the perceived probability of spending too little, about right, or too much to
improve and protect the environment for each year of the GSS from 1973 to 2010,
separately. Survey years during which there was a Republican president in office are
drawn in red; survey years during which there was a Democratic president in office are
drawn in blue. Most individuals think that we are spending too little on the environment
as a society, from about 50% of respondents to 80% of respondents, depending on the
survey year. About 20 to 40% of individuals think that we are spending about right on the
environment. The remaining individuals, about 10%, reported that we are spending too
much to improve and protect the environment. Descriptively, this figure demonstrates
that the perceived probability of spending too little on the environment increases during
Republican presidencies, and conversely, that the perceived probability of spending about
right decreases. Compared to these two larger categories of responses, perceptions of
spending too much do not appear, descriptively, to vary as much over time; the largest
descriptive shift seems to be shifting individuals out of the spending about right category
and into the spending too little category during Republican presidencies.
From the perspective of our policy learning model, the most striking part of this figure is
that the shift in the WTP distributions associated with the political party of the president
seems to accumulate over time. Looking at the longer series of consecutive years of
republican presidency from 1982 to 1991, and from 2002 to 2008, we see that the
perceived probability of spending too little continues to increase over the number of
31
continuous, consecutive years of Republican presidency. For example, the percentage of
respondents who answer that we are spending too little is at about 60% in 1982, 70% in
1987, and nearly 80% in 1990. Similarly, perceptions of spending too little grew from
60% in 2002 to about 65% in 2004 and around 70% in 2006 and 2008. Associations
between consecutive years of Republican presidency and the perceived probability of
spending too much for the environment seem smaller than those for the larger categories
of spending about right and spending too little. While it is true that the absolute change is
smaller for this group, each of the three probability categories is changing from its own
baseline probability, but this interpretation of the results as a relative change is difficult in
figure 1 since each probability is measured on a single, absolute scale from 0 to 1, rather
than as a relative change from its own baseline.
Both theoretically and empirically, it is important to distinguish between mass policy
feedbacks that affect everyone or almost everyone, and partisan bias—that democrats or
liberals may have an adverse reaction to Republican presidents and that Republicans or
conservatives may have a favorable reaction. If the results in Figure 1 represent partisan
bias, then we might expect shifts in the WTP distribution over time associated with
Republican presidency to be concentrated among democrats and liberals; similarly, we
might expect the patterns in Figure 1 to disappear or move in the opposite direction when
examining only those individuals who self-identify as Republicans or conservatives. In
Figures 2 and 3, we repeat the analysis from the first figure, but restricting the estimation
sample to those individuals who consider themselves conservative (either a slight
conservative, conservative, or extreme conservative) in Figure 2, and to those individuals
32
who consider themselves republican (republican or strong republican) in Figure 3.
Importantly, we find the same types of patterns for conservatives and Republicans as we
do for the overall population: that even individuals who consider themselves Republican
or conservative appear to respond to consecutive years of Republican presidency in the
same manner as the overall public. For example, in Figure 3 we see that in 1982, at the
start of a relatively long period of consecutive Republican presidency, about 40% of self-
identified conservatives indicated that they thought that the government was spending too
little to improve and protect the environment. By 1991, after 9 consecutive years of
Republican presidency, about 70% of conservatives had come to think we are spending
too little for the environment. This association was descriptively similar for self-
identified Republicans, who shifted from about 45% in 1982 to nearly 80% in 1991. The
results presented in these figures suggest that the patterns described in Figure 1 may
represent mass feedbacks—public learning in response to the accumulation of policy
failures such as spending cuts over time, rather than being an artifact of partisan bias
applying only to certain subgroups of individuals based on their own personal biases
toward one political party or the other.
A depiction of our main finding of environment policy learning is presented in Figure 4.
Here, we have consecutive years of Republican presidency, from 1 to 10, on the x-axis;
and two y-axes: one for actual spending changes and one for perceptions of the
corresponding spending changes. The first y-axis, on the left-hand side of the graph,
shows the spending change in billions of dollars in 2010 that accumulates over the years
of consecutive Republican administrations, compared to a baseline among democratic
33
presidencies during the study period (the Carter, Clinton, and Obama administrations).
These cumulative spending changes are shown in red bars, compared to a baseline among
democratic presidencies standardized to 0 dollars and demarcated in blue at the top of the
figure. The second y-axis, on the right-hand side of the graph, shows the change in
perceived probability of spending too little to improve and protect the environment that
accumulates over consecutive years of Republican presidency. These perceptions of
spending are plotted for each year of consecutive Republican presidency, 1 through 20, in
black circles (with the size of the circle representing the relative sample size from the
survey). The black line shows the linear fit, simply drawing a line through the perceived
probabilities.
During a Republican presidency, and after adjusting for a nonpartisan, linear time trend
in the WTP distribution, there is a strong linear association between consecutive years of
Republican presidency and the perceived probability of spending too little to protect and
improve the environment. At one year of Republican presidency, less than 60% of
individuals think that we are spending too little for the environment. After 10 years, this
probability increased to nearly 75%, an increase from the year 1 baseline of 25%.
Without examining actual spending changes associated with consecutive years of
Republican presidency, it would not be clear whether or not the survey responses are
accurate; individuals could stereotype Republican presidents as spending too little on the
environment, even if they spent the same, or more, than Democrats. The red bars show
that this is not the case. During the first year of Republican presidency, environmental
spending was cut an average of 5 billion dollars compared to levels during Democratic
34
presidencies. These spending cuts continue while Republicans remain in office, and they
accumulate over time. By 10 years of consecutive Republican presidencies, the
administrations cut inflation-adjusted spending by about 45 billion in 2010 dollars. While
the public responds to Republican administrations over time by providing feedback that
they are thought of as spending too little on the environment, there was no corresponding
change in spending levels by the government to correlate with the feedback; Republicans
continued to cut environment spending, even as the negative feedback against these cuts
continued to increase.
In Figures 5 and 6, we repeat the analysis in Figure 4 for two other policy domains of
interest—health and defense—in order to see whether the feedback effect identified for
the environment represents something particular about the partisan dynamics of
environmental policy, or whether it is some more general pattern that could be found for
other seemingly partisan policy domains. Compared to the environment, feedback for
health and defense spending contained similar directions, but the feedback was noticeably
less in magnitude. We might think that health and environment are both similar partisan
policy domains that are generally supported by individuals who are more likely to
identify as Democrat or liberal, in which case the comparison between the strong
feedback for environment, and the much weaker feedback for health, may seem
confusing or counterintuitive. This pattern of results, however, is consistent with the
literature examining public opinion about health spending, which generally finds the
public demand for government health programs to be surprisingly inelastic for a variety
35
of reasons, at least before the Obama health reforms and vociferous partisan debates in
town hall meetings, media outlets, and other settings.
In the remaining Figures 7 through 11 we address more explicitly the distinction between
mass policy feedbacks among nearly the entirety of the public, as opposed to a partisan
bias, wherein apparent feedbacks might instead indicate a response to consecutive years
of Republican presidency that may only occur for Democrats and liberals, rather than a
mass feedback or pattern of policy learning. When examining the effects of consecutive
years of Republican presidency on the perceived probability of spending too little for the
environment by various subgroups of individuals, our model’s predictions were
consistent with a mass feedback among almost all subgroups of individuals.
Furthermore, the feedbacks were larger for groups of individuals who seem more likely
to be affected by consecutive years of Republican presidency, further supporting our
methodological approach and resulting estimates.
In figure 7, we show again consecutive years of republican presidency from 1 to 10 on
the x-axis and the perceived probability of spending too little to improve and protect the
environment on the y-axis, allowing the linear relationship between these two variables to
vary by the stated political ideology of the individual survey respondent. One year after a
Republican president takes office, and after adjusting for a nonpartisan time trend in the
WTP distribution, as well as individual education, income, and party affiliation, 50% of
conservatives, 60% of moderates, and 70% of liberals think we are spending too little on
the environment. Each of these groups demonstrates policy learning over time, and after
10 years of consecutive Republican presidency, nearly 70% of conservatives, about 73%
36
of moderates, and more than 80% of liberals think that we are spending too little for the
environment as a society. More specific categories within liberal and conservative
ideology groupings are reported in Figures 8 and 9, respectively. At year 1 of a series of
years with Republicans continuously in office, about 67% of slight liberals thought that
we were spending too little for the environment; about 72% of liberals, and the same
percent of extreme liberals, thought that we were spending too little. After 10 years of
continuous, consecutive Republican presidency, all three groups exhibited greater
demand for environmental spending; slight liberals had increased to about 77%, liberals
to 80%, and extreme liberals to more than 90%. Figure 9 shows that slight conservatives,
conservatives, and extreme conservatives all had similar perceptions of environmental
spending at 1 year of Republican presidency—about 50% thought we were spending too
little. By examining the slope of the demand curve over time, we find that it is steepest
for slight conservatives, who increased to about 74%; less steep for conservatives, who
increased to about 67%; and nearly flat for extreme conservatives, who increased by only
1 or 2 percentage points from the year 1 baseline of 50%.
Figures 10 and 11 show the relationship between consecutive years of Republican
presidency, the perceived probability of spending too little to protect and improve the
environment, and party affiliation for coarse and fine groupings of party affiliation,
respectively. When examined by different party affiliations, we see the linear relationship
between consecutive years of Republican presidency and the probability of spending too
little for the environment for each category. In Figure 10, we see that Republicans had a
markedly lower baseline perception of spending being too little at one year of republican
37
presidency than either independents or Democrats -- at one year into a long series of
consecutive years of a Republican administration, about 50% of survey respondents who
themselves were Republican thought that we were spending too little for the
environment, compared to about 62% of Democrats or independents. Each group
increased over time, and at 10 years out, about 67% of Republicans, 72% of
independents, and 75% of Democrats perceived levels of environmental spending as
being too little. Figure 11 shows that, while each group has its own level of demand for
environmental spending, e.g., strong Republicans having the lowest baseline WTP at less
than 50%, the pattern of learning happens for each of the narrower categories of party
affiliation, with each group’s demand curve shifting over time relative to its own
baseline. Even among individuals who consider themselves to be strong Republicans,
demand for environmental spending increased over consecutive years of Republican
presidency, from about 42% to approximately 60%--an increase from their baseline
probability of about 43%.
Conclusion
Based on our results showing that, adjusting for a non-partisan, perceived probability for
spending too little on the environment significantly increases linearly during Republican
presidencies, we can reject our null hypothesis. By measuring actual environmental
spending levels during both Republican and Democratic administrations, we are also able
to determine that the perceived probability for spending too little on the environment
significantly increases (decreases) as actual government appropriations for environmental
protection decrease (increase).
38
The shift over time in the WTP distribution for environmental protection associated with
consecutive years of a political party’s presidency, as well as with cumulative changes in
actual government appropriation levels, is not replicated when analyzing health spending,
which prompts no policy feedback. We do find negligible feedback for defense spending.
These results contradict Wlezien (1995), which empirically shows greater policy
feedback for defense spending than general social (environment, health care, welfare)
spending. The contraction between our results and the literature can be reconciled by our
analysis of specific general social spending items. A potential lack of feedback from
health-specific spending in Wlezien (1995) could have been neutralized by any negative
feedback from environment specific spending. Since Wlezien (1995) does not delineate
specific social spending categories, we cannot confirm this interplay between
environment and health policy feedback.
When distinguishing between mass policy feedback and partisan bias, we noted similar
feedback for various levels of partisan Republicans and Democrats, as well as for various
levels of liberals and conservatives, but failed to observe feedback effects for those who
defined themselves as “extreme conservatives.” This lack of feedback observed from
extreme conservatives may occur as a result of their views on the role of government.
Extreme conservatives might not respond to changes in levels of government spending
for the environment because they believe that it is not the government’s role to provide
environmental protection in the first place, and thus they consistently opposing spending
for it. When first comparing Tables 3 and 4, we might find that our results are
39
contradictory because we find a significant feedback response from “Strong Republicans”
that we do not observe among those who identify themselves as “Extreme
Conservatives.” To understand this perceived discrepancy, we must first realize that
strong Republicans and extreme conservatives are not synonymous. Extreme
conservatives may identify themselves as independents or Libertarians, instead of as
strong Republicans. Since the GSS does not specifically break down the party affiliation
for independents, we cannot confirm the cross affiliation specifically within this dataset.
It is also important to recognize how a lack of environmental knowledge in the electorate
may serve as a barrier to policy feedback. The literature confirms the usage of the
political party of the president as a salient secondary source of information for
environmental protection spending. However, without the proper knowledge and scope of
environmental issues like climate change, Leiserowitz (2010), the electorate may not
know if and/or how much protection is needed from the government to protect against
potential economic and societal damages. This lack of knowledge could disrupt the
feedback process needed to prompt our elected officials on how to properly act upon
environmental issues, such as climate change and pollution.
Our findings imply that the movement or change in opinion for WTP for environmental
protection is driven by both the signal for and actual changes in government
appropriations for environmental protection, and not just by demographics, socio-
economic status, and political affiliation and ideology. Since this phenomenon of policy
feedback is not consistent across other policy domains, we hope to provoke opinion
40
scientists to explore the applied consequences of this relationship specifically in the
context of environmental policy. When predicting support for future environmental
regulations and expenditures, social scientists should consider taking policy feedback into
account in their modeling.
For example, many green and climate policy supporters are still pushing for
comprehensive climate legislation at the state and federal levels that would put a price on
carbon emissions. By almost all estimates, any variation of this legislation that is enacted
will add to public and private expenditures. Therefore, in gauging support for this
legislation, opinion scientists will have to measure the public’s WTP for environmental
protection. Based on our model for policy feedback, and assuming knowledge barriers do
not disrupt the feedback process, we would recommend that these policy advocates
pursue climate legislation in the beginning of a Democratic administration that
immediately followed a Republican one, or at the end of a Republican administration that
is open to increasing appropriations for environmental protection. The latter circumstance
is unlikely, as we note that the government did not push back on electorate feedback by
adjusting appropriation levels to correlate with public opinion.
41
Bibliography
Arcury, Thomas A., and Timothy P. Johnson. “Public Environmental Knowledge: A
Statewide Survey.” The Journal of Environmental Education 18, no. 4 (1987):
:31-37.
Brambor, Thomas, Clark, William R., and Golder, Matt. “Understanding Interaction
Models: Improving Empirical Analyses.” Political Analysis 14, no. 1 (2005): 63-
82.
Carpini, Michael X. Delli, and Scott Keeter. “Measuring Political Knowledge: Putting
First Things First.” American Journal of Political Science 37, no. 4 (1993):1179-
1206.
Converse, Philip E. “The Nature of Belief systems in Mass Publics (1964).” Critical
Review 18, nos. 1-3 (2006): 1-74.
Dunlap, Riley E. “Trends in Public Opinion Toward Environmental Issues: 1965–1990.”
Society & Natural Resources 4, no. 3 (1991): 285-312.
Dunlap, Riley E., and Michael Patrick Allen.“Partisan Differences on Environmental
Issues: A Congressional Roll-Call Analysis.” The Western Political Quarterly 29,
no. 3 (1976): 384-397.
Dunlap, R. E., C. Xiao, and A. M. McCright “Politics and Environment in America:
Partisan and Ideological Cleavages in Public Support for Environmentalism.”
Environmental Politics 10, no. 4 (2001): 23-48.
Durr, Robert H. “What Moves Policy Sentiment?” The American Political Science
Review 87, no. 1 (1993): 158-170.
Eichenberg, Richard C., and Richard Stoll.“Representing Defense: Democratic Control of
42
the Defense Budget in the United States and Western Europe.” The Journal of
Conflict Resolution 47, no. 4 (2003): 399-422.
Elliott, Euel, James L. Regens, and Barry J. Seldon. “Exploring Variation in Public
Support for Environmental Protection.”Social Science Quarterly (University of
Texas Press) 76, no. 1 (1995): 41-52.
Elliott, Euel, Barry J. Seldon, and James L. Regens. “Political and Economic
Determinants of Individuals» Support for Environmental Spending.” Journal of
Environmental Management 51, no. 1 (1997): 15-27.
Erikson, Robert S., Michael B. MacKuen, and James A. Stimson. The Macro Polity. New
York: Cambridge University Press, 2002.
Feldman, Stanley. “Measuring Issue Preferences: The Problem of Response
Instability.”Political Analysis 1, no. 1 (1989): 25-60.
Froot, Kenneth A. “Consistent Covariance Matrix Estimation with Cross-Sectional
Dependence and Heteroskedasticity in Financial Data.” Journal of Financial and
Quantitative Analysis 24, no. 03 (1989): 333-355.
Gelissen, John. “Explaining Popular Support for Environmental Protection.”
Environment and Behavior 39, no. 3 (2007): 392-415.
Gomez, Brad T., and J. Matthew Wilson. “Political Sophistication and Economic Voting
in the American Electorate: A Theory of Heterogeneous Attribution.” American
Journal of Political Science 45, no. 4 (2001): 899-914.
Hanmer, Michael J, and Kerem Ozan Kalkan. “Behind the Curve: Clarifying the Best
Approach to Calculating Predicted Probabilities and Marginal Effects from
Limited Dependent Variable Models.” American Journal of Political Science. doi:
43
10.1111/j.1540-5907.2012.00602.x
Hartley, Thomas, and Bruce Russett. “Public Opinion and the Common Defense: Who
Governs Military Spending in the United States?” The American Political Science
Review 86, no. 4 (1992): 905-915.
Herron, Michael C. “Postestimation Uncertainty in Limited Dependent Variable
Models.” Political Analysis 8, no. 1 (1999): 83-98.
House, US White
2011 Historical Tables. In Table 15.4 -- Total Government Expenditure by
Major Category of Expenditure: 1948-2011. O.o.M.a. Budget, ed.
Hunter, Lori M., and Lesley Rinner. “The Association Between Environmental
Perspective and Knowledge and Concern With Species Diversity.” Society &
Natural Resources 17, no. 6 (2004): 517-532.
Israel, D., Levinson, A. “Willingness to Pay for Environmental Quality: Testable
Empirical Implications of the Growth and Environmental Literature.”
Contributions to Economic Analysis and Policy 3, no. 1 (2004): art. 2. .
Jones, Robert Emmet, and Riley E. Dunlap. “The Social Bases of Environmental
Concern: Have They Changed Over Time?” Rural Sociology 57, no. 1 (1992): 28-
47.
Kohut, A. Doherty, C., Dimock, M., Keeter, S. Pew Research News IQ Quiz: What the
Public Knows about the Political Parties Washington, D.C.: The Pew Research
Center.
Konisky, David M., Jeffrey Milyo, and Lilliard E. Richardson. “Environmental Policy
Attitudes: Issues, Geographical Scale, and Political Trust*.” Social Science
44
Quarterly 89, no. 5 (2008): 1066-1085.
Leiserowitz, A., Smith, N. & Marlon, J.R. Americans’ Knowledge of Climate Change.
New Haven, CT: Yale Project on Climate Change Communication, 2010.
Lorenzoni, Irene, and Nick Pidgeon. “Public Views on Climate Change: European and
USA Perspectives.” Climatic Change 77, no. 1 (2006): 73-95.
McKelvey, Richard D, and William Zavoina. “A Statistical Model for the Analysis of
Ordinal Level Dependent Variables.” The Journal of Mathematical Sociology 4,
no. 1 (1975): 103-120.
Page, Benjamin I., and Robert Y. Shapiro. “Effects of Public Opinion on Policy.” The
American Political Science Review 77, no. 1 (1983): 175-190.
Page, Benjamin I., and Robert Y. Shapiro. The Rational Public: Fifty Years of Trends in
Americans’ Policy Preferences. Chicago: University of Chicago Press, 1992.
Smith, Tom W, et al.“General Social Surveys, 1972-2010 [machine-readable data file].”
Sponsored by National Science Foundation. NORC, ed. Chicago: National
Opinion Research Center. Storrs, CT: The Roper Center for Public Opinion
Research, University of Connecticut, 2011..
Soroka, Stuart N., and Wlezien, Christopher. “Opinion Representation and Policy
Feedback: Canada in Comparative Perspective.” Canadian Journal of Political
Science/Revue Canadienne de Science Politique 37, no. 03 (2004): 531-559.
Soroka, Stuart N., and Wlezien, Christopher. “Opinion–Policy Dynamics: Public
Preferences and Public Expenditure in the United Kingdom.” British Journal of
Political Science 35, no. 04 (2005): 665-689.
Stanley, Harold W., and Niemi, Richard.
45
2009 “Presidents and Vice Presidents of the United States.” In Vital Statistics on
American Politics, 2009-2010, 233-235. Washington, DC: CQ Press.
Stimson, James A. Public Opinion in America: Moods, Cycles, and Swings. Boulder,
Colorado: Westview, 1991.
Stimson, James A., Michael B. MacKuen, and Robert S. Erikson. “Opinion and Policy: A
Global View.” PS: Political Science and Politics 27, no. 1 (1994): 29-35.
Stimson, James A. Tides of Consent. Location: Cambridge University Press, 2004.
Wlezien, Christopher. “The Public as Thermostat: Dynamics of Preferences for
Spending.” American Journal of Political Science 39, no. 4 (1995): 981-1000.
Wlezien, Christopher. “Patterns of Representation: Dynamics of Public Preferences and
Policy.” Journal of Politics 66, no. 1 (2004): 1-24.
Zaller, John, and Stanley Feldman. “A Simple Theory of the Survey Response:
Answering Questions Versus Revealing Preferences.” American Journal of
Political Science 36, no. 3 (1992): 579-616.
46

More Related Content

What's hot

Review social capital and health inequity
Review social capital and health inequityReview social capital and health inequity
Review social capital and health inequity
Veerle Vyncke
 
Democracy and Expertise: Lessons from the Debate over Global Warming
Democracy and Expertise: Lessons from the Debate over Global WarmingDemocracy and Expertise: Lessons from the Debate over Global Warming
Democracy and Expertise: Lessons from the Debate over Global Warming
sebhancock
 
Political Ideological Divides and Actual Views
Political Ideological Divides and Actual ViewsPolitical Ideological Divides and Actual Views
Political Ideological Divides and Actual Views
Michael Silverman
 
Urban Transportation Ecoefficiency: Social and Political Forces for Change in...
Urban Transportation Ecoefficiency: Social and Political Forces for Change in...Urban Transportation Ecoefficiency: Social and Political Forces for Change in...
Urban Transportation Ecoefficiency: Social and Political Forces for Change in...
Anna McCreery
 
Stability in a Changing World: The Social Policy Regimes of Germany, Sweden a...
Stability in a Changing World: The Social Policy Regimes of Germany, Sweden a...Stability in a Changing World: The Social Policy Regimes of Germany, Sweden a...
Stability in a Changing World: The Social Policy Regimes of Germany, Sweden a...
sebhancock
 
Preparing for Genocide: Community Work in Rwanda
Preparing for Genocide: Community Work in RwandaPreparing for Genocide: Community Work in Rwanda
Preparing for Genocide: Community Work in Rwanda
Stockholm Institute of Transition Economics
 
ENST490_finalpaper
ENST490_finalpaperENST490_finalpaper
ENST490_finalpaper
Benjamin Campbell
 
Murray Gabriel Djupe #2
Murray Gabriel Djupe #2Murray Gabriel Djupe #2
Murray Gabriel Djupe #2
Gabriel Murray
 
mitchell_186_final paper copy
mitchell_186_final paper copymitchell_186_final paper copy
mitchell_186_final paper copy
Alec Mitchell
 
AP Public Opinion
AP Public OpinionAP Public Opinion
AP Public Opinion
Taylor Phillips
 
Health Psychology of Urbanicity: Does it Increase Violent Behavior within Ado...
Health Psychology of Urbanicity: Does it Increase Violent Behavior within Ado...Health Psychology of Urbanicity: Does it Increase Violent Behavior within Ado...
Health Psychology of Urbanicity: Does it Increase Violent Behavior within Ado...
Joseph White MPA CPM
 
Moving to Opportunity Experiment
Moving to Opportunity ExperimentMoving to Opportunity Experiment
Moving to Opportunity Experiment
Jonathan Dunnemann
 
2016 Citizen's Committee for Children of New York - Community Risk Ranking
2016 Citizen's Committee for Children of New York - Community Risk Ranking2016 Citizen's Committee for Children of New York - Community Risk Ranking
2016 Citizen's Committee for Children of New York - Community Risk Ranking
Jonathan Dunnemann
 
mitchell_186_final paper copy
mitchell_186_final paper copymitchell_186_final paper copy
mitchell_186_final paper copy
Alec Mitchell
 
Draft3_Davies_PO300
Draft3_Davies_PO300Draft3_Davies_PO300
Draft3_Davies_PO300
Mackenzie Davies
 
RPA Spatial Planning and Inequality Fourth Regional Plan Roundtable
RPA Spatial Planning and Inequality Fourth Regional Plan RoundtableRPA Spatial Planning and Inequality Fourth Regional Plan Roundtable
RPA Spatial Planning and Inequality Fourth Regional Plan Roundtable
Jonathan Dunnemann
 
Nisbet aaas sept2016
Nisbet aaas sept2016Nisbet aaas sept2016
Nisbet aaas sept2016
Matthew Nisbet
 
Consequences of democratic citizens' policy agenda 2
Consequences of democratic citizens' policy agenda 2Consequences of democratic citizens' policy agenda 2
Consequences of democratic citizens' policy agenda 2
jordanlachance
 
USA results
USA resultsUSA results
USA results
Natalie Dyer
 
How research should incorporate gender dimensions to inform climate change po...
How research should incorporate gender dimensions to inform climate change po...How research should incorporate gender dimensions to inform climate change po...
How research should incorporate gender dimensions to inform climate change po...
Gotelind Alber
 

What's hot (20)

Review social capital and health inequity
Review social capital and health inequityReview social capital and health inequity
Review social capital and health inequity
 
Democracy and Expertise: Lessons from the Debate over Global Warming
Democracy and Expertise: Lessons from the Debate over Global WarmingDemocracy and Expertise: Lessons from the Debate over Global Warming
Democracy and Expertise: Lessons from the Debate over Global Warming
 
Political Ideological Divides and Actual Views
Political Ideological Divides and Actual ViewsPolitical Ideological Divides and Actual Views
Political Ideological Divides and Actual Views
 
Urban Transportation Ecoefficiency: Social and Political Forces for Change in...
Urban Transportation Ecoefficiency: Social and Political Forces for Change in...Urban Transportation Ecoefficiency: Social and Political Forces for Change in...
Urban Transportation Ecoefficiency: Social and Political Forces for Change in...
 
Stability in a Changing World: The Social Policy Regimes of Germany, Sweden a...
Stability in a Changing World: The Social Policy Regimes of Germany, Sweden a...Stability in a Changing World: The Social Policy Regimes of Germany, Sweden a...
Stability in a Changing World: The Social Policy Regimes of Germany, Sweden a...
 
Preparing for Genocide: Community Work in Rwanda
Preparing for Genocide: Community Work in RwandaPreparing for Genocide: Community Work in Rwanda
Preparing for Genocide: Community Work in Rwanda
 
ENST490_finalpaper
ENST490_finalpaperENST490_finalpaper
ENST490_finalpaper
 
Murray Gabriel Djupe #2
Murray Gabriel Djupe #2Murray Gabriel Djupe #2
Murray Gabriel Djupe #2
 
mitchell_186_final paper copy
mitchell_186_final paper copymitchell_186_final paper copy
mitchell_186_final paper copy
 
AP Public Opinion
AP Public OpinionAP Public Opinion
AP Public Opinion
 
Health Psychology of Urbanicity: Does it Increase Violent Behavior within Ado...
Health Psychology of Urbanicity: Does it Increase Violent Behavior within Ado...Health Psychology of Urbanicity: Does it Increase Violent Behavior within Ado...
Health Psychology of Urbanicity: Does it Increase Violent Behavior within Ado...
 
Moving to Opportunity Experiment
Moving to Opportunity ExperimentMoving to Opportunity Experiment
Moving to Opportunity Experiment
 
2016 Citizen's Committee for Children of New York - Community Risk Ranking
2016 Citizen's Committee for Children of New York - Community Risk Ranking2016 Citizen's Committee for Children of New York - Community Risk Ranking
2016 Citizen's Committee for Children of New York - Community Risk Ranking
 
mitchell_186_final paper copy
mitchell_186_final paper copymitchell_186_final paper copy
mitchell_186_final paper copy
 
Draft3_Davies_PO300
Draft3_Davies_PO300Draft3_Davies_PO300
Draft3_Davies_PO300
 
RPA Spatial Planning and Inequality Fourth Regional Plan Roundtable
RPA Spatial Planning and Inequality Fourth Regional Plan RoundtableRPA Spatial Planning and Inequality Fourth Regional Plan Roundtable
RPA Spatial Planning and Inequality Fourth Regional Plan Roundtable
 
Nisbet aaas sept2016
Nisbet aaas sept2016Nisbet aaas sept2016
Nisbet aaas sept2016
 
Consequences of democratic citizens' policy agenda 2
Consequences of democratic citizens' policy agenda 2Consequences of democratic citizens' policy agenda 2
Consequences of democratic citizens' policy agenda 2
 
USA results
USA resultsUSA results
USA results
 
How research should incorporate gender dimensions to inform climate change po...
How research should incorporate gender dimensions to inform climate change po...How research should incorporate gender dimensions to inform climate change po...
How research should incorporate gender dimensions to inform climate change po...
 

Viewers also liked

Desain simbol upload
Desain simbol uploadDesain simbol upload
Desain simbol upload
Rijali Cahyo Wicaksono
 
Basic drawing sesi 1
Basic drawing sesi 1Basic drawing sesi 1
Basic drawing sesi 1
Rijali Cahyo Wicaksono
 
мовознавство
мовознавствомовознавство
мовознавствоAvramova_Vera
 
INFORME DE LA FICHA DE AMPLIFICADORES
INFORME DE LA FICHA DE AMPLIFICADORESINFORME DE LA FICHA DE AMPLIFICADORES
INFORME DE LA FICHA DE AMPLIFICADORES
Karliitha Fallaz
 
Sesi 3 kelompok huruf
Sesi 3 kelompok hurufSesi 3 kelompok huruf
Sesi 3 kelompok huruf
Rijali Cahyo Wicaksono
 
Pelajaran 1
Pelajaran 1Pelajaran 1
Pelajaran 1
ela_kholila
 
Медвежий тур
Медвежий турМедвежий тур
Медвежий тур
Егор Ерёмин
 

Viewers also liked (7)

Desain simbol upload
Desain simbol uploadDesain simbol upload
Desain simbol upload
 
Basic drawing sesi 1
Basic drawing sesi 1Basic drawing sesi 1
Basic drawing sesi 1
 
мовознавство
мовознавствомовознавство
мовознавство
 
INFORME DE LA FICHA DE AMPLIFICADORES
INFORME DE LA FICHA DE AMPLIFICADORESINFORME DE LA FICHA DE AMPLIFICADORES
INFORME DE LA FICHA DE AMPLIFICADORES
 
Sesi 3 kelompok huruf
Sesi 3 kelompok hurufSesi 3 kelompok huruf
Sesi 3 kelompok huruf
 
Pelajaran 1
Pelajaran 1Pelajaran 1
Pelajaran 1
 
Медвежий тур
Медвежий турМедвежий тур
Медвежий тур
 

Similar to AP_Thesis_Using_Indirect_Policy_Feedback_12_26_2012

Access provided by University of Redlands (5 May 2016 2346 GM.docx
Access provided by University of Redlands (5 May 2016 2346 GM.docxAccess provided by University of Redlands (5 May 2016 2346 GM.docx
Access provided by University of Redlands (5 May 2016 2346 GM.docx
nettletondevon
 
Social Policy Responsiveness in Developed DemocraciesAu.docx
 Social Policy Responsiveness in Developed DemocraciesAu.docx Social Policy Responsiveness in Developed DemocraciesAu.docx
Social Policy Responsiveness in Developed DemocraciesAu.docx
gertrudebellgrove
 
Education, Intelligence, and Attitude Extremity
Education, Intelligence, and Attitude ExtremityEducation, Intelligence, and Attitude Extremity
Education, Intelligence, and Attitude Extremity
Vishwa Jeet
 
Social Problems, 2016, 63, 284-301 doi 10.1093socprospw00.docx
Social Problems, 2016, 63, 284-301 doi 10.1093socprospw00.docxSocial Problems, 2016, 63, 284-301 doi 10.1093socprospw00.docx
Social Problems, 2016, 63, 284-301 doi 10.1093socprospw00.docx
samuel699872
 
Literature review on policy diffusion
Literature review on policy diffusionLiterature review on policy diffusion
Literature review on policy diffusion
Valerie Niklas
 
Chong-Druckman-FramingTheory Chong-Druckman-FramingTheory
Chong-Druckman-FramingTheory Chong-Druckman-FramingTheoryChong-Druckman-FramingTheory Chong-Druckman-FramingTheory
Chong-Druckman-FramingTheory Chong-Druckman-FramingTheory
SolomonLee15
 
CHAPTER 7The policy processEileen T. O’GradyThere are t
CHAPTER 7The policy processEileen T. O’GradyThere are tCHAPTER 7The policy processEileen T. O’GradyThere are t
CHAPTER 7The policy processEileen T. O’GradyThere are t
JinElias52
 
ENV TERM PAPER
ENV TERM PAPERENV TERM PAPER
ENV TERM PAPER
Kritika Gupta
 
RESEARCH ARTICLETalking about Climate Change and GlobalW.docx
RESEARCH ARTICLETalking about Climate Change and GlobalW.docxRESEARCH ARTICLETalking about Climate Change and GlobalW.docx
RESEARCH ARTICLETalking about Climate Change and GlobalW.docx
debishakespeare
 
Running head HIGHER EDUCATION POLICY1HIGHER EDUCATION POLICY.docx
Running head HIGHER EDUCATION POLICY1HIGHER EDUCATION POLICY.docxRunning head HIGHER EDUCATION POLICY1HIGHER EDUCATION POLICY.docx
Running head HIGHER EDUCATION POLICY1HIGHER EDUCATION POLICY.docx
wlynn1
 
American Public Policy Chapter 1
American Public Policy Chapter 1American Public Policy Chapter 1
American Public Policy Chapter 1
Julie Brown
 
server05productnCCPP5-4CPP403.txt unknown Seq 1 13-OCT.docx
server05productnCCPP5-4CPP403.txt unknown Seq 1 13-OCT.docxserver05productnCCPP5-4CPP403.txt unknown Seq 1 13-OCT.docx
server05productnCCPP5-4CPP403.txt unknown Seq 1 13-OCT.docx
lesleyryder69361
 
Neal D. Buckwalter is assistant profes-sor in the School o.docx
Neal D. Buckwalter is assistant profes-sor in the School o.docxNeal D. Buckwalter is assistant profes-sor in the School o.docx
Neal D. Buckwalter is assistant profes-sor in the School o.docx
hallettfaustina
 
Issue Ownership And Representation A Theory Of Legislative
Issue Ownership And Representation A Theory Of LegislativeIssue Ownership And Representation A Theory Of Legislative
Issue Ownership And Representation A Theory Of Legislative
legal2
 
Converse-McGuire Essay
Converse-McGuire EssayConverse-McGuire Essay
Converse-McGuire Essay
Yuliya Pismennaya
 
Social Work, Politics, and Social Policy Education Applying
Social Work, Politics, and Social Policy Education ApplyingSocial Work, Politics, and Social Policy Education Applying
Social Work, Politics, and Social Policy Education Applying
AlleneMcclendon878
 
CHAPTER 1Incorporating Political Indicatorsinto Comparat.docx
CHAPTER 1Incorporating Political Indicatorsinto Comparat.docxCHAPTER 1Incorporating Political Indicatorsinto Comparat.docx
CHAPTER 1Incorporating Political Indicatorsinto Comparat.docx
walterl4
 
Linking political exposures to child and maternal health outcomes a realist r...
Linking political exposures to child and maternal health outcomes a realist r...Linking political exposures to child and maternal health outcomes a realist r...
Linking political exposures to child and maternal health outcomes a realist r...
Araz Taeihagh
 
Public Opinion and Public Policy.docx
Public Opinion and Public Policy.docxPublic Opinion and Public Policy.docx
Public Opinion and Public Policy.docx
Laxmikant Paudel
 
INTS PAPER FINAL FINAL
INTS PAPER FINAL FINALINTS PAPER FINAL FINAL
INTS PAPER FINAL FINAL
Clayton Daniels
 

Similar to AP_Thesis_Using_Indirect_Policy_Feedback_12_26_2012 (20)

Access provided by University of Redlands (5 May 2016 2346 GM.docx
Access provided by University of Redlands (5 May 2016 2346 GM.docxAccess provided by University of Redlands (5 May 2016 2346 GM.docx
Access provided by University of Redlands (5 May 2016 2346 GM.docx
 
Social Policy Responsiveness in Developed DemocraciesAu.docx
 Social Policy Responsiveness in Developed DemocraciesAu.docx Social Policy Responsiveness in Developed DemocraciesAu.docx
Social Policy Responsiveness in Developed DemocraciesAu.docx
 
Education, Intelligence, and Attitude Extremity
Education, Intelligence, and Attitude ExtremityEducation, Intelligence, and Attitude Extremity
Education, Intelligence, and Attitude Extremity
 
Social Problems, 2016, 63, 284-301 doi 10.1093socprospw00.docx
Social Problems, 2016, 63, 284-301 doi 10.1093socprospw00.docxSocial Problems, 2016, 63, 284-301 doi 10.1093socprospw00.docx
Social Problems, 2016, 63, 284-301 doi 10.1093socprospw00.docx
 
Literature review on policy diffusion
Literature review on policy diffusionLiterature review on policy diffusion
Literature review on policy diffusion
 
Chong-Druckman-FramingTheory Chong-Druckman-FramingTheory
Chong-Druckman-FramingTheory Chong-Druckman-FramingTheoryChong-Druckman-FramingTheory Chong-Druckman-FramingTheory
Chong-Druckman-FramingTheory Chong-Druckman-FramingTheory
 
CHAPTER 7The policy processEileen T. O’GradyThere are t
CHAPTER 7The policy processEileen T. O’GradyThere are tCHAPTER 7The policy processEileen T. O’GradyThere are t
CHAPTER 7The policy processEileen T. O’GradyThere are t
 
ENV TERM PAPER
ENV TERM PAPERENV TERM PAPER
ENV TERM PAPER
 
RESEARCH ARTICLETalking about Climate Change and GlobalW.docx
RESEARCH ARTICLETalking about Climate Change and GlobalW.docxRESEARCH ARTICLETalking about Climate Change and GlobalW.docx
RESEARCH ARTICLETalking about Climate Change and GlobalW.docx
 
Running head HIGHER EDUCATION POLICY1HIGHER EDUCATION POLICY.docx
Running head HIGHER EDUCATION POLICY1HIGHER EDUCATION POLICY.docxRunning head HIGHER EDUCATION POLICY1HIGHER EDUCATION POLICY.docx
Running head HIGHER EDUCATION POLICY1HIGHER EDUCATION POLICY.docx
 
American Public Policy Chapter 1
American Public Policy Chapter 1American Public Policy Chapter 1
American Public Policy Chapter 1
 
server05productnCCPP5-4CPP403.txt unknown Seq 1 13-OCT.docx
server05productnCCPP5-4CPP403.txt unknown Seq 1 13-OCT.docxserver05productnCCPP5-4CPP403.txt unknown Seq 1 13-OCT.docx
server05productnCCPP5-4CPP403.txt unknown Seq 1 13-OCT.docx
 
Neal D. Buckwalter is assistant profes-sor in the School o.docx
Neal D. Buckwalter is assistant profes-sor in the School o.docxNeal D. Buckwalter is assistant profes-sor in the School o.docx
Neal D. Buckwalter is assistant profes-sor in the School o.docx
 
Issue Ownership And Representation A Theory Of Legislative
Issue Ownership And Representation A Theory Of LegislativeIssue Ownership And Representation A Theory Of Legislative
Issue Ownership And Representation A Theory Of Legislative
 
Converse-McGuire Essay
Converse-McGuire EssayConverse-McGuire Essay
Converse-McGuire Essay
 
Social Work, Politics, and Social Policy Education Applying
Social Work, Politics, and Social Policy Education ApplyingSocial Work, Politics, and Social Policy Education Applying
Social Work, Politics, and Social Policy Education Applying
 
CHAPTER 1Incorporating Political Indicatorsinto Comparat.docx
CHAPTER 1Incorporating Political Indicatorsinto Comparat.docxCHAPTER 1Incorporating Political Indicatorsinto Comparat.docx
CHAPTER 1Incorporating Political Indicatorsinto Comparat.docx
 
Linking political exposures to child and maternal health outcomes a realist r...
Linking political exposures to child and maternal health outcomes a realist r...Linking political exposures to child and maternal health outcomes a realist r...
Linking political exposures to child and maternal health outcomes a realist r...
 
Public Opinion and Public Policy.docx
Public Opinion and Public Policy.docxPublic Opinion and Public Policy.docx
Public Opinion and Public Policy.docx
 
INTS PAPER FINAL FINAL
INTS PAPER FINAL FINALINTS PAPER FINAL FINAL
INTS PAPER FINAL FINAL
 

AP_Thesis_Using_Indirect_Policy_Feedback_12_26_2012

  • 1. 1 Using Indirect Policy Feedback to Understand Public Opinion for Improving and Protecting the Environment Jeremy Craig Green Anthony Leiserowitz Arnab Pal Abstract In this study, we measured the extent to which the political party of the President of the United States influences the electorate’s willingness to pay for improving and protecting the environment. We used General Social Survey (GSS) data to develop discrete choice models that dynamically gauged collective public opinion on environmental spending from 1973-2010. This learning model of policy feedback monitors public response to policies as those policies shift over time. Our model identified a strong, linear relationship between consecutive years of Republican presidency and the perceived probability of spending too little to improve and protect the environment. During Democratic administrations, about 57% of individuals consider environmental spending to be too little. This shift accumulates over time, and after 10 years of consecutive Republican administrations, that number rises to 73%. Despite this increased feedback concerning comparably low levels of environmental spending, Republican administrations continued to decrease spending relative to Democrats, and after 10 consecutive years in office, Republicans had decreased spending levels by nearly $40 billion (adjusted for inflation, in 2010 dollars). Individual characteristics (party
  • 2. 2 affiliation, political ideology, income, and education categories) are also assessed using discrete choice methods. Introduction Policy feedback theory holds that as policy actions in one direction or the other increase, public opinion will shift against the given direction of policy, and politicians may then, in turn, respond to opinion by moving their policies in the other direction. This theory requires individuals to have knowledge of these policies and to respond to public opinion surveys accordingly, in order to provide feedback expressing their policy preferences. Politicians, on the other hand, have a responsibility to monitor public opinion and respond to it, although the extent to which this happens continues to be a topic of deep debate among political scientists. This paper provides a fresh perspective on these streams of literature by developing a policy-learning model of public opinion, where responses to public opinion surveys shift over time in association with consecutive years of Republican presidency. We focus on public opinion for willingness to pay for improving and protecting the environment, a domain that has grown increasingly partisan over the past 1-2 decades and importantly, is a topic for which individuals may have relatively little direct, personal experiences with the major policy issues (heat waves might be a noticeably exception), and may thus be likely to base their attitudes on some secondary source of information, such as the political party of the President. Attitudes toward health and military spending by the government are also evaluated in order to compare the extent to which environmental attitudes may be more or less partisan than other policy domains.
  • 3. 3 In this essay, we develop and test a learning model of policy feedback, where the public responds to policies as the policies shift over time. The main innovation of our approach, in comparison to that taken in previous studies, is that we examine more dynamically accumulating changes in public opinion that occur over the course of continuous, consecutive years of Republican presidency. Using data from the General Social Survey (GSS) from 1973 to 2010, discrete choice models were used to test for shifts in the willingness to pay (WTP) distribution for federal spending on the environment and natural resources, as well as for other policy domains. In particular, we tested for a linear relationship between consecutive years of Republican presidency and the perceived probability of spending too little, about right, or too much to improve and protect the environment. To quantify the extent to which individual survey responses represent stereotypes of Republicans as presidents with low levels of environmental spending, or whether Republicans actually do spend less than Democrats over consecutive years of administration, we tested for a corresponding relationship between consecutive years of Republican presidency and actual levels of spending, compared to an empirically robust average of spending during Democratic administrations. Review of the Literature Previous theoretical and empirical descriptions of the relationships between public policy and public opinion offer a variety of explanations for changes in public attitudes over time. Individual attitudes reported in public opinion surveys seem to best reflect changes in public policy if individuals are paying particular attention to a given issue and have
  • 4. 4 meaningful beliefs concerning a given topic, but they might be less meaningful in cases where individuals do not have specific knowledge of a topic at hand. When asked about their preferences for government spending to improve and protect the environment, individuals may not have meaningful beliefs about government spending insofar as they do not know the actual levels of government spending. Although survey respondents may not have access to this primary information about government expenditures, their beliefs might still be meaningful if they can access some type of “secondary” information to help them formulate accurate assessments of a concept on which they have no primary information. Stimson (1991) notes that long-term policy preferences across different issues are connected and refers to these long-term changes as “moods.” Wlezien (1995) describes a negative feedback loop between policy outputs and public inputs that occurs when the public’s preferences for spending on policy are inversely related by a government’s spending decisions. In a Democratic polity, information gaps may be narrowed through the process of negative feedback (Stimson, 2004). According to this theory, if the current policy differs from the desired policy, the public will send a message to policy makers to modify the policy to the public’s preference (i.e., spending level). For example, if a government expenditures increases (decreases) for environment and natural resources spending beyond a desired level (for the public), public support for expenditure on environment and natural resource spending will adjust by decreasing (increasing), suggesting that the public processes information on budgetary policy for these issues accurately.
  • 5. 5 In the case of government spending on environment and natural resources, whether or not public opinion responds to government policy seems to depend on how government policy is conceptualized theoretically and empirically. Some research has suggested that public opinion does not respond to individually delineated social expenditures (such as health, education, environment, cities) by the government (i.e., the information gap persists), which might suggest that the public cannot discriminate specific social policy change with regards to changes in appropriations (Wlezien, 1995). By not responding to government appropriation levels for specific social issues, we could potentially fail to participate in Wlezien’s negative feedback loop with regards to specific social policy related to environmental protection. Importantly, however, even if we are unable to provide an accurate response for how much we are willing to pay for environmental protection through signals created by environmental conditions and/or policy appropriations, we may instead be able to rely on secondary information. Secondary Information According to Page and Shapiro (1992), individuals may adjust their policy preferences based on new policy pertinent facts (i.e., changes in federal appropriations). However, individuals are more likely to respond to this new information “using cognitive shortcuts or rules of thumb, such as reliance upon trusted delegates or reference figures (friends, interest groups, experts, and political leaders) to do the political reasoning for them and to provide guidance.” (Page and Shapiro, 1992) This idea is supported by Gomez and Wilson (2001) in their research regarding the economic judgments of less sophisticated
  • 6. 6 (or knowledgeable) voters, who attribute “responsibilities for outcomes to the most obvious actor in the relevant sphere.” In regard to the national economy, this secondary source would be the President. These less sophisticated voters attribute the national economic condition more so to the President than to other economic forces and policies that cannot be controlled by the President. Gomez and J Wilson (2001) also note that more sophisticated voters are less likely to make this direct attribution and more likely to evaluate the national economic condition based on multiple variables and sources of information. Although secondary information may be more cognitively accessible to the average individual than primary information, it may also be less accurate. Stimson (2004) proposes that the public does receive feedback in terms of how the environment is doing overall (i.e., Is the environment healthy or unhealthy?). Unfortunately, the public’s knowledge of environmental issues has historically been unsophisticated and limited (Arcury, 1987). This gap in the public’s knowledge has persisted with newer environmental concerns, such as climate change (Leiserowitz, 2010), which is an issue that is difficult for the public to understand and relate to in their daily lives (Lorenzoni, 2006). The political party of elected officials might be a particularly salient source of secondary information that individuals might rely on when responding to questions about federal spending on natural resources and the environment. Partisan political leaders have taken positions that coincide with individual voting partisans, as Republicans and conservatives
  • 7. 7 are less likely to support further government efforts to protect the environment (Konisky, D. M., Milyo, J., & Richardson, L., 2008). The history of partisanship in federal environmental spending and legislation dates back to the origins of environmental issues. Environmental legislation has been partisan, as noted by congressional Republicans in the 1970’s, who were significantly less likely than congressional Democrats to vote for the environmental legislation passed by the then-Republican-led Nixon and Ford administrations (Dunlap & Allen, 1976). Within both Congress and the White House, partisan differences in support for environmental protection have only strengthened over time, as Republican congressional leaders and the administrations of Presidents Reagan and George W. Bush have championed anti-environmental protection policy positions, while Democratic administrations and congressional leaders have more homogenously moved towards more pro-environmental protection policy stances (Dunlap, 2001). The political party of the President may serve as the most effective secondary information source for spending on environmental protection. A 2012 PEW Research Center survey noted that more Americans could identify the political party of former presidents Ronald Regan, Bill Clinton, and John F. Kennedy (85%, 84%, 78% respectively) than the political parties of current congressional leaders Nancy Pelosi and John Boehner (61% and 55% respectively) (Kohut et. al). Carpini and Keeter (1993) conducted an analysis of the 1989-1991 National Election Surveys to determine the factual political knowledge of the public by examining various survey questions related to politics and government policy. The surveys asked participants to identify political leaders and their political parties, recognize trends in government appropriations for
  • 8. 8 defense and social policy, and answer various civics questions. An analysis of the surveys showed that the party identification of a former President (Nixon, p=.78) was a more significant predictor of overall factual political knowledge than the current party identification of the congressional (House p=.68 & Senate p=.55) majorities. Both forms of party identification are more significant in predicting political factual knowledge than recognizing trends in government spending for both defense (p=.28) and education spending (p=.24). Public Engagement Other theoretical and empirical literature provides alternative explanations for changes in public opinion over time. It is very difficult for the public to answer a question on how much our government should spend to improve and protect a set of issues, as a proper answer requires knowledge that can only come from full-time attention to a specific issue (Stimson, 2004). The notion that respondents do not have meaningful attitudes or beliefs has been widely contested by academics, e.g., Feldman (1989) & Page and Shapiro (1992). For example, Converse’s (1964) study suggests that large portions of the public may not have meaningful knowledge about survey questions, so they instead offer whatever response they think may please the interviewers. Whether or not individuals are able to provide meaningful responses to a given survey question depends on the question, as well as the policy domain of import. Historically, public knowledge of environmental issues has been limited. Arcury and Johnson (1987) brought the issue of public environmental knowledge to the forefront by
  • 9. 9 using a statewide survey to determine that public knowledge is low and also directly correlates with education, income, and sex. The level of public environmental knowledge has not changed even as environmental issues have. Yale University’s Americans’ Knowledge of Climate Change (2010) survey report noted that over half of all respondents failed (answered less than 60% of 81 graded questions correctly) the public knowledge of climate change survey, thus confirming that a majority of Americans do not have a fully developed understanding of climate change. Hunter and Rinner (2004) explore this public knowledge issue one step further by analyzing the relationship between public concern for and knowledge of environmental issues. The study uses a Colorado state survey that measures public concern for species diversity using the New Ecological Paradigm (NEP) scale, and concludes that those with greater knowledge of the species are not significantly likely to be more concerned with the species itself. Policy Feedback Scholars have long examined associations between public opinion and public policy action from governments (Eichenberg, R. C. and R. Stoll 2003; Erikson, MacKuen, and Stimson 2002). Page and Shapiro (1983) examine public opinion and policy data from 1935 to 1979 to conclude that opinion usually moves before policy and is, thus, an important factor for policy change. When examining public opinion’s impacts on specific public policy (defense spending), Hartley, T. and B. Russett (1992) find evidence that public opinion influences military spending. If public sentiment for more military spending rises, actual government spending for military spending will increase (and vice versa). Stimson, J. A., M. B. MacKuen, et al. (1994) used empirical evidence to support
  • 10. 10 an alternative conclusion, where in the long run, shifts in public opinion are tied to shifts in public policy. Within this relationship where policy is the leading indicator, Wlezien (1995) noted a negative feedback loop in regard to issues of defense, welfare, and general social spending (i.e., appropriations for environment, health, welfare, and big cities). When examining public spending preferences for defense and general social spending between 1977 and 1991, Wlezien (1995) concludes that the public becomes more (less) supportive of government spending as government appropriations decrease (increase). However, the public’s response to military spending is much stronger than its response to social spending. A 1% increase in appropriations for social programs led to a .18% decrease in public support, while an equal increase in defense appropriations led to a decrease in public support (2.7%). An examination of opinion policy dynamics in Canada by Soroka, and Wlezien (2004) note that the public in Canada responded to changes in public spending across various domains, though the response is less pronounced than in the US. Soroka, S. N. and C. Wlezien (2005) conduct a similar analysis of opinion representation and policy feedback in the UK and note through empirical evidence that the public adjusts its preferences for various spending domains in response to spending itself, thereby providing additional evidence for the negative feedback phenomenon present in both the U.S. and Canada. The study notes that the negative feedback effect in the U.K. is more pronounced than the effect in the United States. Wlezien (2004) empirically shows that the United States government adjusts to this negative feedback loop by changing defense, welfare, and general social policy spending to levels that are desired by the public. This policy representation is less pronounced in the U.K. and Canada, where policy makers are
  • 11. 11 less representative of their constituencies under their respective parliamentary systems than they would be under the presidential system present in the United States (Soroka, S. N. and C. Wlezien 2004; Soroka, S. N. and C. Wlezien 2005). Hypothesis Based on Stimson’s theory of negative feedback and our understanding of secondary source information, we expect a positive linear relationship between consecutive years of Republican presidency and the perceived probability of spending too little to improve and protect the environment, even when adjusting for a non-partisan. Null Hypothesis 1: When analyzing the period from 1973-2010 in the GSS survey, the perceived probability of spending too little on the environment in the US electorate, adjusting for a non-partisan, will not significantly and linearly increase with consecutive years of a Republican presidency. Methods Relationships between consecutive years of Republican presidents, willingness to pay for improving and protecting the environment, and individual characteristics (party affiliation, political ideology, and income and education categories) were assessed using discrete choice methods. To test for policy learning—the accumulation of policy feedback over time—we looked for a linear relationship between consecutive years of Republican presidency and willingness to pay for improving and protecting the
  • 12. 12 environment, conditional on a Republican president in office. Choices were modeled using an underlying regression function: (1) The expected WTP for an individual surveyed in year is expressed in equation (1) as a probability function (F) of the number of continuous, consecutive years of Republican presidency, a binary indicator for Republican presidency, and interactions between each of these explanatory variables and individual characteristics: 10 quantiles of family income in constant dollars as approximated by the GSS, 5 education categories, 7 categories of political party affiliations, and 7 categories of political ideology. Substantive implications of the underlying model were assessed by examining the slope of the linear relationship between WTP and consecutive years of Republican presidency with the relative change in WTP from 1 year after a Republican takes office to 10 consecutive years of Republican presidency (9 years after baseline). The slope of the linear relationship may be formally defined as an average marginal effect (AME), or the partial derivatives of the WTP function for each individual , separately, and averaged over the estimation sample, . The relative change from a baseline WTP was estimated from the percent change from a baseline probability one year after a Republican president takes office to 10 consecutive years of Republican presidency (nine years after baseline), conditional on a Republican president in office. Each of these implications of the model was formally tested as follows, respectively:
  • 13. 13 (2) The regression specified in equation (1) was estimated using an ordered probit model (McKelvey and Zavoina 1975)-ERROR; quantities of interest specified in equations (2) and (3) were estimated using the observed-value approach (Hanmer and Ozan Kalkan 2012)-ERROR. Conditional marginal effects were estimated to assess the implicit and explicit interactions in the model (Brambor 2005)-ERROR. Standard errors of interest were clustered on survey year (Froot 1989)-ERROR, and estimated using the delta method (Herron 1999)-ERROR. Data and Study Sample Data on willingness to pay to improve and protect the environment were drawn from the General Social Survey (GSS) (McKelvey and Zavoina 1975; Smith et al. 2011). Most importantly, this survey included a question about individual attitudes toward levels of government spending on the environment from 1973 to 2010 for American adults ages 18 and above: We are faced with many problems in this country, none of which can be solved easily or inexpensively. I'm going to name some of these problems, and for each one I'd like you to name some of these problems, and for each one I'd like you to tell me whether you think we're spending too much money on it, too little money, or about the right amount. Are we spending too much, too little, or about the right amount on improving and protecting the environment? (Smith et al. 2011)
  • 14. 14 From 1973 to 2010, 61% of GSS respondents reported that they believed that our society was spending too little on the environment, 30% thought that spending was about right, and 9% said that we are spending too much. Data on the precise timing of presidencies were pulled from Congressional Quarterly (Hanmer and Ozan Kalkan 2012; Stanley and Niemi 2009). When presidents were both in and out of office during a single survey year, we marked the president who was in office for most of the year. Individual characteristics entered into the model at the highest level of detail available in the GSS and interacted with the key explanatory variables: 7 categories of political ideology and of party affiliation, 5 education categories, and 10 quantiles of family income in constant dollars as approximated by the GSS from the survey categories. Data on government spending for environment and natural resources, health, and arms and military were pulled from the White House Office of Management and Budget (Brambor 2005; US White House 2011). Individuals who did not respond to the question (less than 1%) and those who said that they just “don’t know” (less than 5%) were excluded from the sample, because these two categories do not fit into the ordered levels of WTP responses. Figures for Results
  • 15. 15 0.2.4.6.8 Figure 1: Perceived Probability of Spending to Improve and Protect the Environment, 1973-2010 Pr(Spending Too Little) Pr(Spending About Right) Pr(Spending Too Much) 1970 1980 1990 2000 2010 Survey Year N = 28776
  • 16. 16 0.2.4.6.8 Figure 2: Perceived Probability of Spending to Improve and Protect the Environment, Self-Identified Republicans (1973-2010) Pr(Spending Too Little) Pr(Spending About Right) Pr(Spending Too Much) 1970 1980 1990 2000 2010 Survey Year N = 9894
  • 17. 17 0.2.4.6.8 Figure 3: Perceived Probability of Spending to Improve and Protect the Environment, Self-Identified Conservatives (1974-2010) Pr(Spending Too Little) Pr(Spending About Right) Pr(Spending Too Much) 1970 1980 1990 2000 2010 Survey Year N = 8256
  • 18. 18 -50-40-30-20-100 .6.65.7.75.8 Pr(SpendingTooLittle) Figure 4: Demand for Federal Environment and Natural Resource Spending, by Consecutive Years of Republican Presidency (compared to an average of the Carter, Clinton, and Obama administrations) 0 2 4 6 8 10 Consecutive Years of Republican Presidency N = 27462 Note. Perceived spending is adjusted for a nonpartisan, linear time trend; actual spending is adjusted for inflation.
  • 19. 19 -300-200-1000 .6.65.7.75.8 Pr(SpendingTooLittle) Figure 5: Demand for Federal Health Spending, by Consecutive Years of Republican Presidency (compared to an average of the Carter, Clinton, and Obama administrations) 0 2 4 6 8 10 Consecutive Years of Republican Presidency N = 27462 Note. Perceived spending is adjusted for a nonpartisan, linear time trend; actual spending is adjusted for inflation.
  • 20. 20 02004006008001000 .3.35.4.45.5 Pr(SpendingTooMuch) Figure 6: Demand for Federal Defense Spending, by Consecutive Years of Republican Presidency (compared to an average of the Carter, Clinton, and Obama administrations) 0 2 4 6 8 10 Consecutive Years of Republican Presidency N = 27462 Note. Perceived spending is adjusted for a nonpartisan, linear time trend; actual spending is adjusted for inflation.
  • 21. Figure 7: Public Opinion and Consecutive Years of Republican Presidency, by Political Ideology (1974-2010) 21 Pr(SpendingTooLittle) .5.6.7.8 Self-Identified Political Ideology Liberal Moderate Conservative 1 10 Consecutive Years of Republican Presidency N = 27462 Note. Adjusted for education, income, party affiliation and a nonpartisan, linear time trend.
  • 22. Figure 8: Public Opinion and Consecutive Years of Republican Presidency, by Political Ideology (1974-2010) 22 Pr(SpendingTooLittle) .65.7.75.8.85.9 Self-Identified Political Ideology Extreme Liberal Liberal Slight Liberal 1 10 Consecutive Years of Republican Presidency N = 27462 Note. Adjusted for education, income, party affiliation and a nonpartisan, linear time trend.
  • 23. Figure 9: Public Opinion and Consecutive Years of Republican Presidency, by Political Ideology (1974-2010) 23 Pr(SpendingTooLittle) .5.55.6.65.7.75 Self-Identified Political Ideology Slight Conservative Conservative Extreme Conservative 1 10 Consecutive Years of Republican Presidency N = 27462 Note. Adjusted for education, income, party affiliation and a nonpartisan, linear time trend.
  • 24. Figure 10: Public Opinion and Consecutive Years of Republican Presidency, by Party Affiliation (1974-2010) 24 Pr(SpendingTooLittle) .5.55.6.65.7.75 Self-Identified Party Affiliation Democrat Independent Republican 1 10 Consecutive Years of Republican Presidency N = 27462 Note. Adjusted for education, income, political ideology and a nonpartisan, linear time trend.
  • 25. Figure 11 : Public Opinion and Consecutive Years of Republican Presidency, by Party Affiliation (1974-2010) 25 Pr(SpendingTooLittle) .4.5.6.7.8 Self-Identified Party Affiliation Strong Democrat Democrat Independent Near Democrat Independent Independent Near Republican Republican Strong Republican 1 10 Consecutive Years of Republican Presidency N = 27462 Note. Adjusted for education, income, political ideology and a nonpartisan, linear time trend.
  • 26. 26 Table 1: Associations Between Republican Presidency and Demand for Federal Environment Spending, 1973-2010 1973-2010 Sample 1974-2010 Sample Adjustments: Unadjusted Time Trend Unadjusted Time Trend Covariates Spending Too Little Consecutive Years Republican 0.0160úúú 0.0132úúú 0.0189úúú 0.0162úúú 0.0155úúú (0.00363) (0.00380) (0.00318) (0.00318) (0.00297) % change, 1 to 10 years 23.05úúú 18.92úúú 27.93úúú 23.62úúú 22.25úúú (5.659) (5.740) (5.081) (4.895) (4.630) Spending About Right Consecutive Years Republican -0.00929úúú -0.00769úúú -0.0110úúú -0.00938úúú -0.00905úúú (0.00204) (0.00217) (0.00179) (0.00184) (0.00162) % change, 1 to 10 years -28.19úúú -23.78úúú -32.47úúú -28.29úúú -27.35úúú (5.579) (6.178) (4.748) (5.097) (4.546) Spending Too Much Consecutive Years Republican -0.00669úúú -0.00554úúú -0.00797úúú -0.00681úúú -0.00644úúú (0.00161) (0.00164) (0.00143) (0.00137) (0.00139) % change, 1 to 10 years -51.94úúú -45.35úúú -58.07úúú -52.28úúú -47.13úúú (8.083) (9.640) (6.268) (7.233) (7.728) individuals 28776 28776 27462 27462 27462 survey years 27 27 26 26 26 linear time trend X X X covariate adjustments X Standard errors in parentheses ú p < 0.05, úú p < 0.01, úúú p < 0.001 Note. Covariates are education, income, party affiliation, and political ideology.
  • 27. 27 Table 2: Associations Between Republican Presidency and Demand for Federal Environment, Health, and Defense Spending (1974-2010) Spending Too Little Environment Health Defense Consecutive Years Republican 0.0155úúú 0.00456 -0.0117úú (0.00297) (0.00293) (0.00433) % change, 1 to 10 years 22.25úúú 5.746 -37.18úúú (4.630) (3.870) (11.23) Spending About Right Consecutive Years Republican -0.00905úúú -0.00305 -0.00181 (0.00162) (0.00195) (0.00111) % change, 1 to 10 years -27.35úúú -10.48 -7.837ú (4.546) (6.323) (3.461) Spending Too Much Consecutive Years Republican -0.00644úúú -0.00152 0.0135úú (0.00139) (0.00102) (0.00514) % change, 1 to 10 years -47.13úúú -19.06 40.40ú (7.728) (12.36) (17.56) individuals = 27462, survey years = 26 ú p < 0.05, úú p < 0.01, úúú p < 0.001 Note. Adjusted for education, income, party affiliation, political ideology and a nonpartisan, linear time trend.
  • 28. 28 Table 3: By ideology Extreme Liberal Liberal Slight Liberal Moderate Slight Conservative Conservative Extreme Conservative Spending Too Little Consecutive Years Republican 0.0290úúú 0.0101 0.0128úú 0.0169úúú 0.0264úúú 0.0199úúú 0.000885 (0.00576) (0.00644) (0.00393) (0.00379) (0.00511) (0.00571) (0.00779) % change, 1 to 10 years 28.65úúú 11.47 15.91úú 24.23úúú 43.55úúú 35.29úúú 1.555 (5.113) (7.346) (4.975) (5.919) (10.32) (10.26) (13.62) Spending About Right Consecutive Years Republican -0.0209úúú -0.00728 -0.00869úúú -0.0101úúú -0.0136úúú -0.00955úúú -0.000544 (0.00375) (0.00449) (0.00248) (0.00202) (0.00193) (0.00274) (0.00374) % change, 1 to 10 years -67.92úúú -26.65 -28.54úúú -29.77úúú -39.12úúú -27.18úúú -2.128 (8.624) (15.30) (7.500) (5.642) (5.968) (8.168) (9.389) Spending Too Much Consecutive Years Republican -0.00810úúú -0.00277 -0.00412úú -0.00677úúú -0.0128úúú -0.0103úúú -0.000341 (0.00206) (0.00196) (0.00146) (0.00181) (0.00327) (0.00302) (0.00406) % change, 1 to 10 years -85.61úúú -41.65 -46.40úúú -51.66úúú -66.73úúú -53.27úúú -0.281 (6.582) (24.20) (11.66) (8.978) (7.769) (11.68) (27.95) individuals = 27462, survey years = 26 ú p < 0.05, úú p < 0.01, úúú p < 0.001 Note. Adjusted for education, income, party affiliation and a nonpartisan, linear time trend.
  • 29. 29 Table 4: By party Strong Democrat Democrat Independent Near Democrat Independent Independent Near Republican Republican Strong Republican Spending Too Little Consecutive Years Republican 0.0176úúú 0.0138úúú 0.0105úú 0.0156úúú 0.0143ú 0.0201úúú 0.0172ú (0.00344) (0.00289) (0.00324) (0.00424) (0.00605) (0.00345) (0.00696) % change, 1 to 10 years 22.61úúú 18.89úúú 12.99úú 21.32úúú 21.73ú 31.76úúú 34.44ú (4.826) (4.269) (4.223) (5.726) (9.539) (6.079) (15.54) Spending About Right Consecutive Years Republican -0.0113úúú -0.00846úúú -0.00707úúú -0.00961úúú -0.00805ú -0.0108úúú -0.00693úú (0.00191) (0.00157) (0.00204) (0.00250) (0.00316) (0.00164) (0.00252) % change, 1 to 10 years -35.11úúú -25.91úúú -23.42úúú -29.34úúú -23.77úú -31.51úúú -20.76ú (5.377) (4.547) (6.203) (7.386) (9.171) (4.695) (8.099) Spending Too Much Consecutive Years Republican -0.00632úúú -0.00530úúú -0.00347úú -0.00602úúú -0.00627ú -0.00935úúú -0.0102ú (0.00156) (0.00134) (0.00122) (0.00176) (0.00293) (0.00188) (0.00454) % change, 1 to 10 years -54.72úúú -42.80úúú -36.23úúú -47.93úúú -41.85úú -54.65úúú -44.48úú (7.888) (7.681) (10.88) (10.82) (15.62) (6.799) (14.69) individuals = 27462, survey years = 26 ú p < 0.05, úú p < 0.01, úúú p < 0.001 Note. Adjusted for education, income, political ideology and a nonpartisan, linear time trend.
  • 30. 30 Results Descriptive statistics of the estimation sample are displayed in Figure 1. This figure shows the perceived probability of spending too little, about right, or too much to improve and protect the environment for each year of the GSS from 1973 to 2010, separately. Survey years during which there was a Republican president in office are drawn in red; survey years during which there was a Democratic president in office are drawn in blue. Most individuals think that we are spending too little on the environment as a society, from about 50% of respondents to 80% of respondents, depending on the survey year. About 20 to 40% of individuals think that we are spending about right on the environment. The remaining individuals, about 10%, reported that we are spending too much to improve and protect the environment. Descriptively, this figure demonstrates that the perceived probability of spending too little on the environment increases during Republican presidencies, and conversely, that the perceived probability of spending about right decreases. Compared to these two larger categories of responses, perceptions of spending too much do not appear, descriptively, to vary as much over time; the largest descriptive shift seems to be shifting individuals out of the spending about right category and into the spending too little category during Republican presidencies. From the perspective of our policy learning model, the most striking part of this figure is that the shift in the WTP distributions associated with the political party of the president seems to accumulate over time. Looking at the longer series of consecutive years of republican presidency from 1982 to 1991, and from 2002 to 2008, we see that the perceived probability of spending too little continues to increase over the number of
  • 31. 31 continuous, consecutive years of Republican presidency. For example, the percentage of respondents who answer that we are spending too little is at about 60% in 1982, 70% in 1987, and nearly 80% in 1990. Similarly, perceptions of spending too little grew from 60% in 2002 to about 65% in 2004 and around 70% in 2006 and 2008. Associations between consecutive years of Republican presidency and the perceived probability of spending too much for the environment seem smaller than those for the larger categories of spending about right and spending too little. While it is true that the absolute change is smaller for this group, each of the three probability categories is changing from its own baseline probability, but this interpretation of the results as a relative change is difficult in figure 1 since each probability is measured on a single, absolute scale from 0 to 1, rather than as a relative change from its own baseline. Both theoretically and empirically, it is important to distinguish between mass policy feedbacks that affect everyone or almost everyone, and partisan bias—that democrats or liberals may have an adverse reaction to Republican presidents and that Republicans or conservatives may have a favorable reaction. If the results in Figure 1 represent partisan bias, then we might expect shifts in the WTP distribution over time associated with Republican presidency to be concentrated among democrats and liberals; similarly, we might expect the patterns in Figure 1 to disappear or move in the opposite direction when examining only those individuals who self-identify as Republicans or conservatives. In Figures 2 and 3, we repeat the analysis from the first figure, but restricting the estimation sample to those individuals who consider themselves conservative (either a slight conservative, conservative, or extreme conservative) in Figure 2, and to those individuals
  • 32. 32 who consider themselves republican (republican or strong republican) in Figure 3. Importantly, we find the same types of patterns for conservatives and Republicans as we do for the overall population: that even individuals who consider themselves Republican or conservative appear to respond to consecutive years of Republican presidency in the same manner as the overall public. For example, in Figure 3 we see that in 1982, at the start of a relatively long period of consecutive Republican presidency, about 40% of self- identified conservatives indicated that they thought that the government was spending too little to improve and protect the environment. By 1991, after 9 consecutive years of Republican presidency, about 70% of conservatives had come to think we are spending too little for the environment. This association was descriptively similar for self- identified Republicans, who shifted from about 45% in 1982 to nearly 80% in 1991. The results presented in these figures suggest that the patterns described in Figure 1 may represent mass feedbacks—public learning in response to the accumulation of policy failures such as spending cuts over time, rather than being an artifact of partisan bias applying only to certain subgroups of individuals based on their own personal biases toward one political party or the other. A depiction of our main finding of environment policy learning is presented in Figure 4. Here, we have consecutive years of Republican presidency, from 1 to 10, on the x-axis; and two y-axes: one for actual spending changes and one for perceptions of the corresponding spending changes. The first y-axis, on the left-hand side of the graph, shows the spending change in billions of dollars in 2010 that accumulates over the years of consecutive Republican administrations, compared to a baseline among democratic
  • 33. 33 presidencies during the study period (the Carter, Clinton, and Obama administrations). These cumulative spending changes are shown in red bars, compared to a baseline among democratic presidencies standardized to 0 dollars and demarcated in blue at the top of the figure. The second y-axis, on the right-hand side of the graph, shows the change in perceived probability of spending too little to improve and protect the environment that accumulates over consecutive years of Republican presidency. These perceptions of spending are plotted for each year of consecutive Republican presidency, 1 through 20, in black circles (with the size of the circle representing the relative sample size from the survey). The black line shows the linear fit, simply drawing a line through the perceived probabilities. During a Republican presidency, and after adjusting for a nonpartisan, linear time trend in the WTP distribution, there is a strong linear association between consecutive years of Republican presidency and the perceived probability of spending too little to protect and improve the environment. At one year of Republican presidency, less than 60% of individuals think that we are spending too little for the environment. After 10 years, this probability increased to nearly 75%, an increase from the year 1 baseline of 25%. Without examining actual spending changes associated with consecutive years of Republican presidency, it would not be clear whether or not the survey responses are accurate; individuals could stereotype Republican presidents as spending too little on the environment, even if they spent the same, or more, than Democrats. The red bars show that this is not the case. During the first year of Republican presidency, environmental spending was cut an average of 5 billion dollars compared to levels during Democratic
  • 34. 34 presidencies. These spending cuts continue while Republicans remain in office, and they accumulate over time. By 10 years of consecutive Republican presidencies, the administrations cut inflation-adjusted spending by about 45 billion in 2010 dollars. While the public responds to Republican administrations over time by providing feedback that they are thought of as spending too little on the environment, there was no corresponding change in spending levels by the government to correlate with the feedback; Republicans continued to cut environment spending, even as the negative feedback against these cuts continued to increase. In Figures 5 and 6, we repeat the analysis in Figure 4 for two other policy domains of interest—health and defense—in order to see whether the feedback effect identified for the environment represents something particular about the partisan dynamics of environmental policy, or whether it is some more general pattern that could be found for other seemingly partisan policy domains. Compared to the environment, feedback for health and defense spending contained similar directions, but the feedback was noticeably less in magnitude. We might think that health and environment are both similar partisan policy domains that are generally supported by individuals who are more likely to identify as Democrat or liberal, in which case the comparison between the strong feedback for environment, and the much weaker feedback for health, may seem confusing or counterintuitive. This pattern of results, however, is consistent with the literature examining public opinion about health spending, which generally finds the public demand for government health programs to be surprisingly inelastic for a variety
  • 35. 35 of reasons, at least before the Obama health reforms and vociferous partisan debates in town hall meetings, media outlets, and other settings. In the remaining Figures 7 through 11 we address more explicitly the distinction between mass policy feedbacks among nearly the entirety of the public, as opposed to a partisan bias, wherein apparent feedbacks might instead indicate a response to consecutive years of Republican presidency that may only occur for Democrats and liberals, rather than a mass feedback or pattern of policy learning. When examining the effects of consecutive years of Republican presidency on the perceived probability of spending too little for the environment by various subgroups of individuals, our model’s predictions were consistent with a mass feedback among almost all subgroups of individuals. Furthermore, the feedbacks were larger for groups of individuals who seem more likely to be affected by consecutive years of Republican presidency, further supporting our methodological approach and resulting estimates. In figure 7, we show again consecutive years of republican presidency from 1 to 10 on the x-axis and the perceived probability of spending too little to improve and protect the environment on the y-axis, allowing the linear relationship between these two variables to vary by the stated political ideology of the individual survey respondent. One year after a Republican president takes office, and after adjusting for a nonpartisan time trend in the WTP distribution, as well as individual education, income, and party affiliation, 50% of conservatives, 60% of moderates, and 70% of liberals think we are spending too little on the environment. Each of these groups demonstrates policy learning over time, and after 10 years of consecutive Republican presidency, nearly 70% of conservatives, about 73%
  • 36. 36 of moderates, and more than 80% of liberals think that we are spending too little for the environment as a society. More specific categories within liberal and conservative ideology groupings are reported in Figures 8 and 9, respectively. At year 1 of a series of years with Republicans continuously in office, about 67% of slight liberals thought that we were spending too little for the environment; about 72% of liberals, and the same percent of extreme liberals, thought that we were spending too little. After 10 years of continuous, consecutive Republican presidency, all three groups exhibited greater demand for environmental spending; slight liberals had increased to about 77%, liberals to 80%, and extreme liberals to more than 90%. Figure 9 shows that slight conservatives, conservatives, and extreme conservatives all had similar perceptions of environmental spending at 1 year of Republican presidency—about 50% thought we were spending too little. By examining the slope of the demand curve over time, we find that it is steepest for slight conservatives, who increased to about 74%; less steep for conservatives, who increased to about 67%; and nearly flat for extreme conservatives, who increased by only 1 or 2 percentage points from the year 1 baseline of 50%. Figures 10 and 11 show the relationship between consecutive years of Republican presidency, the perceived probability of spending too little to protect and improve the environment, and party affiliation for coarse and fine groupings of party affiliation, respectively. When examined by different party affiliations, we see the linear relationship between consecutive years of Republican presidency and the probability of spending too little for the environment for each category. In Figure 10, we see that Republicans had a markedly lower baseline perception of spending being too little at one year of republican
  • 37. 37 presidency than either independents or Democrats -- at one year into a long series of consecutive years of a Republican administration, about 50% of survey respondents who themselves were Republican thought that we were spending too little for the environment, compared to about 62% of Democrats or independents. Each group increased over time, and at 10 years out, about 67% of Republicans, 72% of independents, and 75% of Democrats perceived levels of environmental spending as being too little. Figure 11 shows that, while each group has its own level of demand for environmental spending, e.g., strong Republicans having the lowest baseline WTP at less than 50%, the pattern of learning happens for each of the narrower categories of party affiliation, with each group’s demand curve shifting over time relative to its own baseline. Even among individuals who consider themselves to be strong Republicans, demand for environmental spending increased over consecutive years of Republican presidency, from about 42% to approximately 60%--an increase from their baseline probability of about 43%. Conclusion Based on our results showing that, adjusting for a non-partisan, perceived probability for spending too little on the environment significantly increases linearly during Republican presidencies, we can reject our null hypothesis. By measuring actual environmental spending levels during both Republican and Democratic administrations, we are also able to determine that the perceived probability for spending too little on the environment significantly increases (decreases) as actual government appropriations for environmental protection decrease (increase).
  • 38. 38 The shift over time in the WTP distribution for environmental protection associated with consecutive years of a political party’s presidency, as well as with cumulative changes in actual government appropriation levels, is not replicated when analyzing health spending, which prompts no policy feedback. We do find negligible feedback for defense spending. These results contradict Wlezien (1995), which empirically shows greater policy feedback for defense spending than general social (environment, health care, welfare) spending. The contraction between our results and the literature can be reconciled by our analysis of specific general social spending items. A potential lack of feedback from health-specific spending in Wlezien (1995) could have been neutralized by any negative feedback from environment specific spending. Since Wlezien (1995) does not delineate specific social spending categories, we cannot confirm this interplay between environment and health policy feedback. When distinguishing between mass policy feedback and partisan bias, we noted similar feedback for various levels of partisan Republicans and Democrats, as well as for various levels of liberals and conservatives, but failed to observe feedback effects for those who defined themselves as “extreme conservatives.” This lack of feedback observed from extreme conservatives may occur as a result of their views on the role of government. Extreme conservatives might not respond to changes in levels of government spending for the environment because they believe that it is not the government’s role to provide environmental protection in the first place, and thus they consistently opposing spending for it. When first comparing Tables 3 and 4, we might find that our results are
  • 39. 39 contradictory because we find a significant feedback response from “Strong Republicans” that we do not observe among those who identify themselves as “Extreme Conservatives.” To understand this perceived discrepancy, we must first realize that strong Republicans and extreme conservatives are not synonymous. Extreme conservatives may identify themselves as independents or Libertarians, instead of as strong Republicans. Since the GSS does not specifically break down the party affiliation for independents, we cannot confirm the cross affiliation specifically within this dataset. It is also important to recognize how a lack of environmental knowledge in the electorate may serve as a barrier to policy feedback. The literature confirms the usage of the political party of the president as a salient secondary source of information for environmental protection spending. However, without the proper knowledge and scope of environmental issues like climate change, Leiserowitz (2010), the electorate may not know if and/or how much protection is needed from the government to protect against potential economic and societal damages. This lack of knowledge could disrupt the feedback process needed to prompt our elected officials on how to properly act upon environmental issues, such as climate change and pollution. Our findings imply that the movement or change in opinion for WTP for environmental protection is driven by both the signal for and actual changes in government appropriations for environmental protection, and not just by demographics, socio- economic status, and political affiliation and ideology. Since this phenomenon of policy feedback is not consistent across other policy domains, we hope to provoke opinion
  • 40. 40 scientists to explore the applied consequences of this relationship specifically in the context of environmental policy. When predicting support for future environmental regulations and expenditures, social scientists should consider taking policy feedback into account in their modeling. For example, many green and climate policy supporters are still pushing for comprehensive climate legislation at the state and federal levels that would put a price on carbon emissions. By almost all estimates, any variation of this legislation that is enacted will add to public and private expenditures. Therefore, in gauging support for this legislation, opinion scientists will have to measure the public’s WTP for environmental protection. Based on our model for policy feedback, and assuming knowledge barriers do not disrupt the feedback process, we would recommend that these policy advocates pursue climate legislation in the beginning of a Democratic administration that immediately followed a Republican one, or at the end of a Republican administration that is open to increasing appropriations for environmental protection. The latter circumstance is unlikely, as we note that the government did not push back on electorate feedback by adjusting appropriation levels to correlate with public opinion.
  • 41. 41 Bibliography Arcury, Thomas A., and Timothy P. Johnson. “Public Environmental Knowledge: A Statewide Survey.” The Journal of Environmental Education 18, no. 4 (1987): :31-37. Brambor, Thomas, Clark, William R., and Golder, Matt. “Understanding Interaction Models: Improving Empirical Analyses.” Political Analysis 14, no. 1 (2005): 63- 82. Carpini, Michael X. Delli, and Scott Keeter. “Measuring Political Knowledge: Putting First Things First.” American Journal of Political Science 37, no. 4 (1993):1179- 1206. Converse, Philip E. “The Nature of Belief systems in Mass Publics (1964).” Critical Review 18, nos. 1-3 (2006): 1-74. Dunlap, Riley E. “Trends in Public Opinion Toward Environmental Issues: 1965–1990.” Society & Natural Resources 4, no. 3 (1991): 285-312. Dunlap, Riley E., and Michael Patrick Allen.“Partisan Differences on Environmental Issues: A Congressional Roll-Call Analysis.” The Western Political Quarterly 29, no. 3 (1976): 384-397. Dunlap, R. E., C. Xiao, and A. M. McCright “Politics and Environment in America: Partisan and Ideological Cleavages in Public Support for Environmentalism.” Environmental Politics 10, no. 4 (2001): 23-48. Durr, Robert H. “What Moves Policy Sentiment?” The American Political Science Review 87, no. 1 (1993): 158-170. Eichenberg, Richard C., and Richard Stoll.“Representing Defense: Democratic Control of
  • 42. 42 the Defense Budget in the United States and Western Europe.” The Journal of Conflict Resolution 47, no. 4 (2003): 399-422. Elliott, Euel, James L. Regens, and Barry J. Seldon. “Exploring Variation in Public Support for Environmental Protection.”Social Science Quarterly (University of Texas Press) 76, no. 1 (1995): 41-52. Elliott, Euel, Barry J. Seldon, and James L. Regens. “Political and Economic Determinants of Individuals» Support for Environmental Spending.” Journal of Environmental Management 51, no. 1 (1997): 15-27. Erikson, Robert S., Michael B. MacKuen, and James A. Stimson. The Macro Polity. New York: Cambridge University Press, 2002. Feldman, Stanley. “Measuring Issue Preferences: The Problem of Response Instability.”Political Analysis 1, no. 1 (1989): 25-60. Froot, Kenneth A. “Consistent Covariance Matrix Estimation with Cross-Sectional Dependence and Heteroskedasticity in Financial Data.” Journal of Financial and Quantitative Analysis 24, no. 03 (1989): 333-355. Gelissen, John. “Explaining Popular Support for Environmental Protection.” Environment and Behavior 39, no. 3 (2007): 392-415. Gomez, Brad T., and J. Matthew Wilson. “Political Sophistication and Economic Voting in the American Electorate: A Theory of Heterogeneous Attribution.” American Journal of Political Science 45, no. 4 (2001): 899-914. Hanmer, Michael J, and Kerem Ozan Kalkan. “Behind the Curve: Clarifying the Best Approach to Calculating Predicted Probabilities and Marginal Effects from Limited Dependent Variable Models.” American Journal of Political Science. doi:
  • 43. 43 10.1111/j.1540-5907.2012.00602.x Hartley, Thomas, and Bruce Russett. “Public Opinion and the Common Defense: Who Governs Military Spending in the United States?” The American Political Science Review 86, no. 4 (1992): 905-915. Herron, Michael C. “Postestimation Uncertainty in Limited Dependent Variable Models.” Political Analysis 8, no. 1 (1999): 83-98. House, US White 2011 Historical Tables. In Table 15.4 -- Total Government Expenditure by Major Category of Expenditure: 1948-2011. O.o.M.a. Budget, ed. Hunter, Lori M., and Lesley Rinner. “The Association Between Environmental Perspective and Knowledge and Concern With Species Diversity.” Society & Natural Resources 17, no. 6 (2004): 517-532. Israel, D., Levinson, A. “Willingness to Pay for Environmental Quality: Testable Empirical Implications of the Growth and Environmental Literature.” Contributions to Economic Analysis and Policy 3, no. 1 (2004): art. 2. . Jones, Robert Emmet, and Riley E. Dunlap. “The Social Bases of Environmental Concern: Have They Changed Over Time?” Rural Sociology 57, no. 1 (1992): 28- 47. Kohut, A. Doherty, C., Dimock, M., Keeter, S. Pew Research News IQ Quiz: What the Public Knows about the Political Parties Washington, D.C.: The Pew Research Center. Konisky, David M., Jeffrey Milyo, and Lilliard E. Richardson. “Environmental Policy Attitudes: Issues, Geographical Scale, and Political Trust*.” Social Science
  • 44. 44 Quarterly 89, no. 5 (2008): 1066-1085. Leiserowitz, A., Smith, N. & Marlon, J.R. Americans’ Knowledge of Climate Change. New Haven, CT: Yale Project on Climate Change Communication, 2010. Lorenzoni, Irene, and Nick Pidgeon. “Public Views on Climate Change: European and USA Perspectives.” Climatic Change 77, no. 1 (2006): 73-95. McKelvey, Richard D, and William Zavoina. “A Statistical Model for the Analysis of Ordinal Level Dependent Variables.” The Journal of Mathematical Sociology 4, no. 1 (1975): 103-120. Page, Benjamin I., and Robert Y. Shapiro. “Effects of Public Opinion on Policy.” The American Political Science Review 77, no. 1 (1983): 175-190. Page, Benjamin I., and Robert Y. Shapiro. The Rational Public: Fifty Years of Trends in Americans’ Policy Preferences. Chicago: University of Chicago Press, 1992. Smith, Tom W, et al.“General Social Surveys, 1972-2010 [machine-readable data file].” Sponsored by National Science Foundation. NORC, ed. Chicago: National Opinion Research Center. Storrs, CT: The Roper Center for Public Opinion Research, University of Connecticut, 2011.. Soroka, Stuart N., and Wlezien, Christopher. “Opinion Representation and Policy Feedback: Canada in Comparative Perspective.” Canadian Journal of Political Science/Revue Canadienne de Science Politique 37, no. 03 (2004): 531-559. Soroka, Stuart N., and Wlezien, Christopher. “Opinion–Policy Dynamics: Public Preferences and Public Expenditure in the United Kingdom.” British Journal of Political Science 35, no. 04 (2005): 665-689. Stanley, Harold W., and Niemi, Richard.
  • 45. 45 2009 “Presidents and Vice Presidents of the United States.” In Vital Statistics on American Politics, 2009-2010, 233-235. Washington, DC: CQ Press. Stimson, James A. Public Opinion in America: Moods, Cycles, and Swings. Boulder, Colorado: Westview, 1991. Stimson, James A., Michael B. MacKuen, and Robert S. Erikson. “Opinion and Policy: A Global View.” PS: Political Science and Politics 27, no. 1 (1994): 29-35. Stimson, James A. Tides of Consent. Location: Cambridge University Press, 2004. Wlezien, Christopher. “The Public as Thermostat: Dynamics of Preferences for Spending.” American Journal of Political Science 39, no. 4 (1995): 981-1000. Wlezien, Christopher. “Patterns of Representation: Dynamics of Public Preferences and Policy.” Journal of Politics 66, no. 1 (2004): 1-24. Zaller, John, and Stanley Feldman. “A Simple Theory of the Survey Response: Answering Questions Versus Revealing Preferences.” American Journal of Political Science 36, no. 3 (1992): 579-616.
  • 46. 46