Are Canadian Political Parties Responsive to Climate Change’s Public Salience in Question Period?
In a new working paper, R. Michael Alvarez and I study issue responsiveness in Canadian politics. Specifically, we consider whether political parties are responsive to their constituents’ policy priorities in choosing the topics they publicly address. We focus our analysis on two of the most pressing policy issues of our time: the environment and climate change.
We use novel measures of issue attention and salience to conduct our analysis. In particular, we quantify the attention political parties pay to the various issues through the topic composition of the speeches emanating from their members during Question Period and policy issues’ public salience by their popularity on Google’s search engine. We believe that both measures are representative of the attention and importance politicians and members of the public attribute to policy issues.
Question Period is a focal moment in Canadian politics. This 45-minute segment occurs every day the House of Commons sits. Its purpose is to offer members of Parliament an opportunity to seek information on the issues of the day and hold the government accountable for its decisions. We employ topic models to reduce the dimension of inherently high-dimensional transcripts and uncover their latent topic composition.
The environment and climate change remain disputed issues, with discernible partisan heterogeneity in attitudes toward them. We expect the latter to result in partisan heterogeneity in issue responsiveness.
In particular, we expect issue responsiveness to vary along two dimensions. First, since parties do not wish to draw attention to problems over which they have a weak reputation since it would weaken their position, they selectively emphasize the issues over which they have a strong(er) reputation and neglect those over which they have a weak(er) reputation. Accordingly, we expect the Conservatives to be less responsive to changes in the environment and climate change’s public salience than the Liberals and New Democrats. Second, through sustained inquiries, opposition parties may compel the government to address policy issues it would otherwise disregard because they are either disadvantageous or embarrassing. The government will ultimately give in to some of the opposition’s pressures and address compromising topics. Accordingly, we expect the Conservatives to be more responsive to the environment and climate change’s public salience in government than in the opposition.
The previous literature has shown that there usually is a strong correlation between the public’s priorities and their political representatives’ agenda.
The figure below corroborates this stylized fact: political parties talk more about the environment and climate during the Question Period when these policy issues are more salient to the public.
For context, this figure depicts, on the y-axis, the log ratio of a party’s speeches dedicated to the environment and climate change in a given week and, on the x-axis, the log-transformed public salience of these issues in Canada over the same period. The relationship between both variables is essentially linear, and its slope is practically indistinguishable between the parties.
Even though previous results show that there is a positive correlation between policy issues’ salience and the topic composition of parties’ speeches, we cannot affirm at this point that the parties alter their actions in reaction to changes in their constituents’ priorities, let alone provide a causal estimate of parties’ issue responsiveness.
Two mechanisms can justify the relationship between the public’s policy priorities and political representatives’ attention. On the one hand, politicians have strong incentives to alter their rhetoric to address chiefly the policy issues salient to their constituents. On the other hand, politicians can engage in “public agenda-setting” and influence the weight voters attribute to the various policy dimensions. Reality lies between these two conflicting theories: even if voters’ sense of priorities is susceptible to manipulation, politicians cannot control all the relevant factors, and exogenous events alter policy issues’ salience. It follows that the relationship between politicians’ attention to policy issues and their public salience is afflicted by simultaneous causality.
By naively observing the correlation between the topic composition of politicians’ public statements and issues’ public salience, we might be under the illusion that politicians are being responsive to their constituents when, instead, citizens’ sense of priorities is being manipulated by politicians’ rhetoric.
We implement a two-stage least squares (2SLS) estimation strategy to remedy this problem. In particular, we instrument the environment and climate change’s public salience in Canada with the analogous variable for the United States. This identification strategy allows us to disentangle the effect of public salience on parties’ speeches (i.e., what we want to measure) from the effect of speeches on salience.
The following table presents estimates of parties’ issue responsiveness. The table contains ordinary least squares (OLS) and 2SLS estimates of four model specifications. These specifications vary along two dimensions: (i) whether partisan heterogeneity in issue responsiveness is permitted or not, and (ii) whether issue responsiveness is allowed to differ after the 2015 election when the Liberal Party replaced the Conservative Party in government.
We highlight three key results. First, all parties are responsive and alter their statements’ topic composition in reaction to exogenous changes in the environment and climate change’s public salience. As seen in column (5), the ratio of a party’s speeches related to the environment and climate change increases on average by 0.388 % following a one percent increase in these issues’ public salience.
Second, two-stage least squares estimation reveals significant heterogeneity in issue responsiveness across the three parties.
Unexpectedly, the Liberal Party is less responsive to changes in the environment and climate change’s public salience than the Conservative Party and the NDP. However, this does not imply that the Liberals talk less about the environment and climate change. Indeed, Liberals seem actually to talk about these issues on their terms. This can be seen by comparing columns (2) and (6): estimation via instrumental variables lowers the estimate of the Liberals’ issue responsiveness by 25 %. This suggests that the Liberals had successfully bolstered the environment and climate change’s public salience over our study period, especially when they were in government.
Third, the Conservative Party and the New Democratic Party have the same estimated level of issue responsiveness. This is remarkable because the environment and climate change have historically been weaker policy issues for the Conservatives. The fact that the extent to which they react to changes in the environment and climate change’s public salience is not significantly different from the New Democrats’ gives credence to the power of Question Period in pushing the party in government to address embarrassing or otherwise unfavorable policy issues.
In addition to parties’ issue responsiveness, we evaluate whether changes in the environment and climate change’s public salience affect the topic composition of Question Period speeches across other topics. We have found evidence that the Conservative Party engaged in obfuscation in reaction to changes in the environment and climate change’s public salience. For instance, the Conservatives obfuscate attention paid to climate change by putting forth economic considerations. Besides, following an increase in the environment and climate change’s public salience, the Conservative Party significantly increases the prevalence of intergovernmental affairs in its Question Period speeches. Since federalism is one of the fundamental principles over which the CPC was founded, this result seems a priori consistent with its ideological matrix. That said, it might also indicate a systematic attempt to minimize the federal government’s responsibility in addressing climate change and pass it on to the provinces and territories instead.
In conclusion, we make notable contributions to the study of political representation and issue responsiveness in Canadian politics. We demonstrated the power of the Question Period in delivering democratic accountability. By attributing agenda-setting power to opposition parties, Question Period allows them to compel the party in government to address issues salient to the public that it may otherwise avoid or neglect. We also uncovered evidence that, during its time in government, the Liberal Party has successfully bolstered the environment and climate change’s public salience through its rhetoric. This result stresses the parties’ role in shaping the political agenda and drawing attention to emerging policy concerns such as the climate crisis. Finally, from a methodological perspective, this paper exemplifies how machine learning can be combined with standard causal identification strategies to answer substantive questions in political science. The methodological innovations we have introduced will considerably benefit the literature on political representation and issue responsiveness.
